1 systemd System and Service Manager
5 * DNSSEC is now turned on by default in systemd-resolved (in
6 "allow-downgrade" mode), but may be turned off during compile time by
7 passing "--with-default-dnssec=no" to "configure" (and of course,
8 during runtime with DNSSEC= in resolved.conf). We recommend
9 downstreams to leave this on at least during development cycles and
10 report any issues with the DNSSEC logic upstream. We are very
11 interested in collecting feedback about the DNSSEC validator and its
12 limitations in the wild. Note however, that DNSSEC support is
13 probably nothing downstreams should turn on in stable distros just
14 yet, as it might create incompatibilities with a few DNS servers and
15 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
16 automatically whenever we detect such incompatible setups, but there
17 might be systems we do not cover yet. Hence: please help us testing
18 the DNSSEC code, leave this on where you can, report back, but then
19 again don't consider turning this on in your stable, LTS or
20 production release just yet. (Note that you have to enable
21 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
22 and its DNSSEC mode for host name resolution from local
25 * systemd-resolve conveniently resolves DANE records with the --tlsa
26 option and OPENPGPKEY records with the --openpgp option. It also
27 supports dumping raw DNS record data via the new --raw= switch.
29 * systemd-logind will now by default terminate user processes that are
30 part of the user session scope unit (session-XX.scope) when the user
31 logs out. This behavior is controlled by the KillUserProcesses=
32 setting in logind.conf, and the previous default of "no" is now
33 changed to "yes". This means that user sessions will be properly
34 cleaned up after, but additional steps are necessary to allow
35 intentionally long-running processes to survive logout.
37 While the user is logged in at least once, user@.service is running,
38 and any service that should survive the end of any individual login
39 session can be started at a user service or scope using systemd-run.
40 systemd-run(1) man page has been extended with an example which shows
41 how to run screen in a scope unit underneath user@.service. The same
42 command works for tmux.
44 After the user logs out of all sessions, user@.service will be
45 terminated too, by default, unless the user has "lingering" enabled.
46 To effectively allow users to run long-term tasks even if they are
47 logged out, lingering must be enabled for them. See loginctl(1) for
48 details. The default polkit policy was modified to allow users to
49 set lingering for themselves without authentication.
51 Previous defaults can be restored at compile time by the
52 --without-kill-user-processes option to "configure".
54 * systemd-logind gained new configuration settings SessionsMax= and
55 InhibitorsMax=, both with a default of 8192. It will not register new
56 user sessions or inhibitors above this limit.
58 * systemd-logind will now reload configuration on SIGHUP.
60 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
61 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
62 enable. Also, support for the "io" cgroup controller in the unified
63 hierarchy has been added, so that the "memory", "pids" and "io" are
64 now the controllers that are supported on the unified hierarchy.
66 WARNING: it is not possible to use previous systemd versions with
67 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
68 is necessary to also update systemd in the initramfs if using the
69 unified hierarchy. An updated SELinux policy is also required.
71 * LLDP support has been extended, and both passive (receive-only) and
72 active (sender) modes are supported. Passive mode ("routers-only") is
73 enabled by default in systemd-networkd. Active LLDP mode is enabled
74 by default for containers on the internal network. The "networkctl
75 lldp" command may be used to list information gathered. "networkctl
76 status" will also show basic LLDP information on connected peers now.
78 * The IAID and DUID unique identifier sent in DHCP requests may now be
79 configured for the system and each .network file managed by
80 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
82 * systemd-networkd gained support for configuring proxy ARP support for
83 each interface, via the ProxyArp= setting in .network files. It also
84 gained support for configuring the multicast querier feature of
85 bridge devices, via the new MulticastQuerier= setting in .netdev
86 files. Similarly, snooping on the IGMP traffic can be controlled
87 via the new setting MulticastSnooping=.
89 A new setting PreferredLifetime= has been added for addresses
90 configured in .network file to configure the lifetime intended for an
93 The systemd-networkd DHCP server gained the option EmitRouter=, which
94 defaults to yes, to configure whether the DHCP Option 3 (Router)
97 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
98 systemd-socket-activate and installed into /usr/bin. It is now fully
101 * systemd-journald now uses separate threads to flush changes to disk
102 when closing journal files, thus reducing impact of slow disk I/O on
105 * The sd-journal API gained two new calls
106 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
107 can be used to open journal files using file descriptors instead of
108 file or directory paths. sd_journal_open_container() has been
109 deprecated, sd_journal_open_directory_fd() should be used instead
110 with the flag SD_JOURNAL_OS_ROOT.
112 * journalctl learned a new output mode "-o short-unix" that outputs log
113 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
114 UTC). It also gained support for a new --no-hostname setting to
115 suppress the hostname column in the family of "short" output modes.
117 * systemd-ask-password now optionally skips printing of the password to
118 stdout with --no-output which can be useful in scripts.
120 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
121 (devices tagged with ID_MAKER_TOOL) are now tagged with
122 "uaccess" and are available to logged in users.
124 * The DeviceAllow= unit setting now supports specifiers (with "%").
126 * "systemctl show" gained a new --value switch, which allows print a
127 only the contents of a specific unit property, without also printing
128 the property's name. Similar support was added to "show*" verbs
129 of loginctl and machinectl that output "key=value" lists.
131 * A new unit type "generated" was added for files dynamically generated
132 by generator tools. Similarly, a new unit type "transient" is used
133 for unit files created using the runtime API. "systemctl enable" will
134 refuse to operate on such files.
136 * A new command "systemctl revert" has been added that may be used to
137 revert to the vendor version of a unit file, in case local changes
138 have been made by adding drop-ins or overriding the unit file.
140 * "machinectl clean" gained a new verb to automatically remove all or
141 just hidden container images.
143 * systemd-tmpfiles gained support for a new line type "e" for emptying
144 directories, if they exist, without creating them if they don't.
146 * systemd-nspawn gained support for automatically patching the UID/GIDs
147 of the owners and the ACLs of all files and directories in a
148 container tree to match the UID/GID user namespacing range selected
149 for the container invocation. This mode is enabled via the new
150 --private-user-chown switch. It also gained support for automatically
151 choosing a free, previously unused UID/GID range when starting a
152 container, via the new --private-users=pick setting (which implies
153 --private-user-chown). Together, these options for the first time
154 make user namespacing for nspawn containers fully automatic and thus
155 deployable. The systemd-nspaw@.service template unit file has been
156 changed to use this functionality by default.
158 * systemd-nspawn gained a new --network-zone= switch, that allows
159 creating ad-hoc virtual Ethernet links between multiple containers,
160 that only exist as long as at least one container referencing them is
161 running. This allows easy connecting of multiple containers with a
162 common link that implements an Ethernet broadcast domain. Each of
163 these network "zones" may be named relatively freely by the user, and
164 may be referenced by any number of containers, but each container may
165 only reference one of these "zones". On the lower level, this is
166 implemented by an automatically managed bridge network interface for
167 each zone, that is created when the first container referencing its
168 zone is created and removed when the last one referencing its zone
171 * The default start timeout may now be configured on the kernel command
172 line via systemd.default_timeout_start_sec=. It was already
173 configurable via the DefaultTimeoutStartSec= option in
174 /etc/systemd/system.conf.
176 * Socket units gained a new TriggerLimitIntervalSec= and
177 TriggerLimitBurst= setting to configure a limit on the activation
178 rate of the socket unit.
180 * The LimitNICE= setting now optionally takes normal UNIX nice values
181 in addition to the raw integer limit value. If the specified
182 parameter is prefixed with "+" or "-" and is in the range -20..19 the
183 value is understood as UNIX nice value. If not prefixed like this it
184 is understood as raw RLIMIT_NICE limit.
186 * Note that the effect of the PrivateDevices= unit file setting changed
187 slightly with this release: the per-device /dev file system will be
188 mounted read-only from this version on, and will have "noexec"
189 set. This (minor) change of behavior might cause some (exceptional)
190 legacy software to break, when PrivateDevices=yes is set for its
191 service. Please leave PrivateDevices= off if you run into problems
194 * systemd-bootchart has been split out to a separate repository:
195 https://github.com/systemd/systemd-bootchart
197 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
198 merged into the kernel in its current form.
200 * The compatibility libraries libsystemd-daemon.so,
201 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
202 which have been deprecated since systemd-209 have been removed along
203 with the corresponding pkg-config files. All symbols provided by
204 those libraries are provided by libsystemd.so.
206 * The Capabilities= unit file setting has been removed (it is ignored
207 for backwards compatibility). AmbientCapabilities= and
208 CapabilityBoundingSet= should be used instead.
210 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
211 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
212 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
213 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
214 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
215 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
216 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
217 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
218 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
219 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
220 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
221 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
222 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
223 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
224 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
225 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
226 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
227 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
228 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
229 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
230 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
231 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
232 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
233 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
234 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
237 — Fairfax, 2016-05-21
241 * The systemd-resolved DNS resolver service has gained a substantial
242 set of new features, most prominently it may now act as a DNSSEC
243 validating stub resolver. DNSSEC mode is currently turned off by
244 default, but is expected to be turned on by default in one of the
245 next releases. For now, we invite everybody to test the DNSSEC logic
246 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
247 service also gained a full set of D-Bus interfaces, including calls
248 to configure DNS and DNSSEC settings per link (for use by external
249 network management software). systemd-resolved and systemd-networkd
250 now distinguish between "search" and "routing" domains. The former
251 are used to qualify single-label names, the latter are used purely
252 for routing lookups within certain domains to specific links.
253 resolved now also synthesizes RRs for all entries from /etc/hosts.
255 * The systemd-resolve tool (which is a client utility for
256 systemd-resolved) has been improved considerably and is now fully
257 supported and documented. Hence it has moved from /usr/lib/systemd to
260 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
263 * The coredump collection logic has been reworked: when a coredump is
264 collected it is now written to disk, compressed and processed
265 (including stacktrace extraction) from a new instantiated service
266 systemd-coredump@.service, instead of directly from the
267 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
268 processing large coredumps can take up a substantial amount of
269 resources and time, and this previously happened entirely outside of
270 systemd's service supervision. With the new logic the core_pattern
271 hook only does minimal metadata collection before passing off control
272 to the new instantiated service, which is configured with a time
273 limit, a nice level and other settings to minimize negative impact on
274 the rest of the system. Also note that the new logic will honour the
275 RLIMIT_CORE setting of the crashed process, which now allows users
276 and processes to turn off coredumping for their processes by setting
279 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
280 and all forked processes by default. Previously, PID 1 would leave
281 the setting at "0" for all processes, as set by the kernel. Note that
282 the resource limit traditionally has no effect on the generated
283 coredumps on the system if the /proc/sys/kernel/core_pattern hook
284 logic is used. Since the limit is now honoured (see above) its
285 default has been changed so that the coredumping logic is enabled by
286 default for all processes, while allowing specific opt-out.
288 * When the stacktrace is extracted from processes of system users, this
289 is now done as "systemd-coredump" user, in order to sandbox this
290 potentially security sensitive parsing operation. (Note that when
291 processing coredumps of normal users this is done under the user ID
292 of process that crashed, as before.) Packagers should take notice
293 that it is now necessary to create the "systemd-coredump" system user
294 and group at package installation time.
296 * The systemd-activate socket activation testing tool gained support
297 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
298 and --seqpacket switches. It also has been extended to support both
299 new-style and inetd-style file descriptor passing. Use the new
300 --inetd switch to request inetd-style file descriptor passing.
302 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
303 variable, which takes a boolean value. If set to false, ANSI color
304 output is disabled in the tools even when run on a terminal that
307 * The VXLAN support in networkd now supports two new settings
308 DestinationPort= and PortRange=.
310 * A new systemd.machine_id= kernel command line switch has been added,
311 that may be used to set the machine ID in /etc/machine-id if it is
312 not initialized yet. This command line option has no effect if the
313 file is already initialized.
315 * systemd-nspawn gained a new --as-pid2 switch that invokes any
316 specified command line as PID 2 rather than PID 1 in the
317 container. In this mode PID 1 is a minimal stub init process that
318 implements the special POSIX and Linux semantics of PID 1 regarding
319 signal and child process management. Note that this stub init process
320 is implemented in nspawn itself and requires no support from the
321 container image. This new logic is useful to support running
322 arbitrary commands in the container, as normal processes are
323 generally not prepared to run as PID 1.
325 * systemd-nspawn gained a new --chdir= switch for setting the current
326 working directory for the process started in the container.
328 * "journalctl /dev/sda" will now output all kernel log messages for
329 specified device from the current boot, in addition to all devices
330 that are parents of it. This should make log output about devices
331 pretty useful, as long as kernel drivers attach enough metadata to
332 the log messages. (The usual SATA drivers do.)
334 * The sd-journal API gained two new calls
335 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
336 that report whether log data from /run or /var has been found.
338 * journalctl gained a new switch "--fields" that prints all journal
339 record field names currently in use in the journal. This is backed
340 by two new sd-journal API calls sd_journal_enumerate_fields() and
341 sd_journal_restart_fields().
343 * Most configurable timeouts in systemd now expect an argument of
344 "infinity" to turn them off, instead of "0" as before. The semantics
345 from now on is that a timeout of "0" means "now", and "infinity"
346 means "never". To maintain backwards compatibility, "0" continues to
347 turn off previously existing timeout settings.
349 * "systemctl reload-or-try-restart" has been renamed to "systemctl
350 try-reload-or-restart" to clarify what it actually does: the "try"
351 logic applies to both reloading and restarting, not just restarting.
352 The old name continues to be accepted for compatibility.
354 * On boot-up, when PID 1 detects that the system clock is behind the
355 release date of the systemd version in use, the clock is now set
356 to the latter. Previously, this was already done in timesyncd, in order
357 to avoid running with clocks set to the various clock epochs such as
358 1902, 1938 or 1970. With this change the logic is now done in PID 1
359 in addition to timesyncd during early boot-up, so that it is enforced
360 before the first process is spawned by systemd. Note that the logic
361 in timesyncd remains, as it is more comprehensive and ensures
362 clock monotonicity by maintaining a persistent timestamp file in
363 /var. Since /var is generally not available in earliest boot or the
364 initrd, this part of the logic remains in timesyncd, and is not done
367 * Support for tweaking details in net_cls.class_id through the
368 NetClass= configuration directive has been removed, as the kernel
369 people have decided to deprecate that controller in cgroup v2.
370 Userspace tools such as nftables are moving over to setting rules
371 that are specific to the full cgroup path of a task, which obsoletes
372 these controllers anyway. The NetClass= directive is kept around for
373 legacy compatibility reasons. For a more in-depth description of the
374 kernel change, please refer to the respective upstream commit:
376 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
378 * A new service setting RuntimeMaxSec= has been added that may be used
379 to specify a maximum runtime for a service. If the timeout is hit, the
380 service is terminated and put into a failure state.
382 * A new service setting AmbientCapabilities= has been added. It allows
383 configuration of additional Linux process capabilities that are
384 passed to the activated processes. This is only available on very
387 * The process resource limit settings in service units may now be used
388 to configure hard and soft limits individually.
390 * The various libsystemd APIs such as sd-bus or sd-event now publicly
391 expose support for gcc's __attribute__((cleanup())) C extension.
392 Specifically, for many object destructor functions alternative
393 versions have been added that have names suffixed with "p" and take a
394 pointer to a pointer to the object to destroy, instead of just a
395 pointer to the object itself. This is useful because these destructor
396 functions may be used directly as parameters to the cleanup
397 construct. Internally, systemd has been a heavy user of this GCC
398 extension for a long time, and with this change similar support is
399 now available to consumers of the library outside of systemd. Note
400 that by using this extension in your sources compatibility with old
401 and strictly ANSI compatible C compilers is lost. However, all gcc or
402 LLVM versions of recent years support this extension.
404 * Timer units gained support for a new setting RandomizedDelaySec= that
405 allows configuring some additional randomized delay to the configured
406 time. This is useful to spread out timer events to avoid load peaks in
407 clusters or larger setups.
409 * Calendar time specifications now support sub-second accuracy.
411 * Socket units now support listening on SCTP and UDP-lite protocol
414 * The sd-event API now comes with a full set of man pages.
416 * Older versions of systemd contained experimental support for
417 compressing journal files and coredumps with the LZ4 compressor that
418 was not compatible with the lz4 binary (due to API limitations of the
419 lz4 library). This support has been removed; only support for files
420 compatible with the lz4 binary remains. This LZ4 logic is now
421 officially supported and no longer considered experimental.
423 * The dkr image import logic has been removed again from importd. dkr's
424 micro-services focus doesn't fit into the machine image focus of
425 importd, and quickly got out of date with the upstream dkr API.
427 * Creation of the /run/lock/lockdev/ directory was dropped from
428 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
429 been available for many years. If you still need this, you need to
430 create your own tmpfiles.d config file with:
432 d /run/lock/lockdev 0775 root lock -
434 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
435 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
436 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
437 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
438 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
439 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
440 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
441 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
442 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
443 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
444 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
445 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
446 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
447 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
448 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
449 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
450 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
451 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
452 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
458 * A number of properties previously only settable in unit
459 files are now also available as properties to set when
460 creating transient units programmatically via the bus, as it
461 is exposed with systemd-run's --property=
462 setting. Specifically, these are: SyslogIdentifier=,
463 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
464 EnvironmentFile=, ReadWriteDirectories=,
465 ReadOnlyDirectories=, InaccessibleDirectories=,
466 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
468 * When creating transient services via the bus API it is now
469 possible to pass in a set of file descriptors to use as
470 STDIN/STDOUT/STDERR for the invoked process.
472 * Slice units may now be created transiently via the bus APIs,
473 similar to the way service and scope units may already be
476 * Wherever systemd expects a calendar timestamp specification
477 (like in journalctl's --since= and --until= switches) UTC
478 timestamps are now supported. Timestamps suffixed with "UTC"
479 are now considered to be in Universal Time Coordinated
480 instead of the local timezone. Also, timestamps may now
481 optionally be specified with sub-second accuracy. Both of
482 these additions also apply to recurring calendar event
483 specification, such as OnCalendar= in timer units.
485 * journalctl gained a new "--sync" switch that asks the
486 journal daemon to write all so far unwritten log messages to
487 disk and sync the files, before returning.
489 * systemd-tmpfiles learned two new line types "q" and "Q" that
490 operate like "v", but also set up a basic btrfs quota
491 hierarchy when used on a btrfs file system with quota
494 * tmpfiles' "v", "q" and "Q" will now create a plain directory
495 instead of a subvolume (even on a btrfs file system) if the
496 root directory is a plain directory, and not a
497 subvolume. This should simplify things with certain chroot()
498 environments which are not aware of the concept of btrfs
501 * systemd-detect-virt gained a new --chroot switch to detect
502 whether execution takes place in a chroot() environment.
504 * CPUAffinity= now takes CPU index ranges in addition to
507 * The various memory-related resource limit settings (such as
508 LimitAS=) now understand the usual K, M, G, ... suffixes to
509 the base of 1024 (IEC). Similar, the time-related resource
510 limit settings understand the usual min, h, day, ...
513 * There's a new system.conf setting DefaultTasksMax= to
514 control the default TasksMax= setting for services and
515 scopes running on the system. (TasksMax= is the primary
516 setting that exposes the "pids" cgroup controller on systemd
517 and was introduced in the previous systemd release.) The
518 setting now defaults to 512, which means services that are
519 not explicitly configured otherwise will only be able to
520 create 512 processes or threads at maximum, from this
521 version on. Note that this means that thread- or
522 process-heavy services might need to be reconfigured to set
523 TasksMax= to a higher value. It is sufficient to set
524 TasksMax= in these specific unit files to a higher value, or
525 even "infinity". Similar, there's now a logind.conf setting
526 UserTasksMax= that defaults to 4096 and limits the total
527 number of processes or tasks each user may own
528 concurrently. nspawn containers also have the TasksMax=
529 value set by default now, to 8192. Note that all of this
530 only has an effect if the "pids" cgroup controller is
531 enabled in the kernel. The general benefit of these changes
532 should be a more robust and safer system, that provides a
533 certain amount of per-service fork() bomb protection.
535 * systemd-nspawn gained the new --network-veth-extra= switch
536 to define additional and arbitrarily-named virtual Ethernet
537 links between the host and the container.
539 * A new service execution setting PassEnvironment= has been
540 added that allows importing select environment variables
541 from PID1's environment block into the environment block of
544 * Timer units gained support for a new RemainAfterElapse=
545 setting which takes a boolean argument. It defaults on on,
546 exposing behaviour unchanged to previous releases. If set to
547 off, timer units are unloaded after they elapsed if they
548 cannot elapse again. This is particularly useful for
549 transient timer units, which shall not stay around longer
550 than until they first elapse.
552 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
553 default now (the kernel default is 16). This is beneficial
554 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
555 allows substantially larger numbers of queued
556 datagrams. This should increase the capability of systemd to
557 parallelize boot-up, as logging and sd_notify() are unlikely
558 to stall execution anymore. If you need to change the value
559 from the new defaults, use the usual sysctl.d/ snippets.
561 * The compression framing format used by the journal or
562 coredump processing has changed to be in line with what the
563 official LZ4 tools generate. LZ4 compression support in
564 systemd was considered unsupported previously, as the format
565 was not compatible with the normal tools. With this release
566 this has changed now, and it is hence safe for downstream
567 distributions to turn it on. While not compressing as well
568 as the XZ, LZ4 is substantially faster, which makes
569 it a good default choice for the compression logic in the
570 journal and in coredump handling.
572 * Any reference to /etc/mtab has been dropped from
573 systemd. The file has been obsolete since a while, but
574 systemd refused to work on systems where it was incorrectly
575 set up (it should be a symlink or non-existent). Please make
576 sure to update to util-linux 2.27.1 or newer in conjunction
577 with this systemd release, which also drops any reference to
578 /etc/mtab. If you maintain a distribution make sure that no
579 software you package still references it, as this is a
580 likely source of bugs. There's also a glibc bug pending,
581 asking for removal of any reference to this obsolete file:
583 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
585 Note that only util-linux versions built with
586 --enable-libmount-force-mountinfo are supported.
588 * Support for the ".snapshot" unit type has been removed. This
589 feature turned out to be little useful and little used, and
590 has now been removed from the core and from systemctl.
592 * The dependency types RequiresOverridable= and
593 RequisiteOverridable= have been removed from systemd. They
594 have been used only very sparingly to our knowledge and
595 other options that provide a similar effect (such as
596 systemctl --mode=ignore-dependencies) are much more useful
597 and commonly used. Moreover, they were only half-way
598 implemented as the option to control behaviour regarding
599 these dependencies was never added to systemctl. By removing
600 these dependency types the execution engine becomes a bit
601 simpler. Unit files that use these dependencies should be
602 changed to use the non-Overridable dependency types
603 instead. In fact, when parsing unit files with these
604 options, that's what systemd will automatically convert them
605 too, but it will also warn, asking users to fix the unit
606 files accordingly. Removal of these dependency types should
607 only affect a negligible number of unit files in the wild.
609 * Behaviour of networkd's IPForward= option changed
610 (again). It will no longer maintain a per-interface setting,
611 but propagate one way from interfaces where this is enabled
612 to the global kernel setting. The global setting will be
613 enabled when requested by a network that is set up, but
614 never be disabled again. This change was made to make sure
615 IPv4 and IPv6 behaviour regarding packet forwarding is
616 similar (as the Linux IPv6 stack does not support
617 per-interface control of this setting) and to minimize
620 * In unit files the behaviour of %u, %U, %h, %s has
621 changed. These specifiers will now unconditionally resolve
622 to the various user database fields of the user that the
623 systemd instance is running as, instead of the user
624 configured in the specific unit via User=. Note that this
625 effectively doesn't change much, as resolving of these
626 specifiers was already turned off in the --system instance
627 of systemd, as we cannot do NSS lookups from PID 1. In the
628 --user instance of systemd these specifiers where correctly
629 resolved, but hardly made any sense, since the user instance
630 lacks privileges to do user switches anyway, and User= is
631 hence useless. Morever, even in the --user instance of
632 systemd behaviour was awkward as it would only take settings
633 from User= assignment placed before the specifier into
634 account. In order to unify and simplify the logic around
635 this the specifiers will now always resolve to the
636 credentials of the user invoking the manager (which in case
637 of PID 1 is the root user).
639 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
640 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
641 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
642 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
643 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
644 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
645 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
646 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
647 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
648 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
649 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
650 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
651 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
652 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
659 * systemd now depends on util-linux v2.27. More specifically,
660 the newly added mount monitor feature in libmount now
661 replaces systemd's former own implementation.
663 * libmount mandates /etc/mtab not to be regular file, and
664 systemd now enforces this condition at early boot.
665 /etc/mtab has been deprecated and warned about for a very
666 long time, so systems running systemd should already have
667 stopped having this file around as anything else than a
668 symlink to /proc/self/mounts.
670 * Support for the "pids" cgroup controller has been added. It
671 allows accounting the number of tasks in a cgroup and
672 enforcing limits on it. This adds two new setting
673 TasksAccounting= and TasksMax= to each unit, as well as a
674 global option DefaultTasksAccounting=.
676 * Support for the "net_cls" cgroup controller has been added.
677 It allows assigning a net class ID to each task in the
678 cgroup, which can then be used in firewall rules and traffic
679 shaping configurations. Note that the kernel netfilter net
680 class code does not currently work reliably for ingress
681 packets on unestablished sockets.
683 This adds a new config directive called NetClass= to CGroup
684 enabled units. Allowed values are positive numbers for fixed
685 assignments and "auto" for picking a free value
688 * 'systemctl is-system-running' now returns 'offline' if the
689 system is not booted with systemd. This command can now be
690 used as a substitute for 'systemd-notify --booted'.
692 * Watchdog timeouts have been increased to 3 minutes for all
693 in-tree service files. Apparently, disk IO issues are more
694 frequent than we hoped, and user reported >1 minute waiting
697 * 'machine-id-commit' functionality has been merged into
698 'machine-id-setup --commit'. The separate binary has been
701 * The WorkingDirectory= directive in unit files may now be set
702 to the special value '~'. In this case, the working
703 directory is set to the home directory of the user
706 * "machinectl shell" will now open the shell in the home
707 directory of the selected user by default.
709 * The CrashChVT= configuration file setting is renamed to
710 CrashChangeVT=, following our usual logic of not
711 abbreviating unnecessarily. The old directive is still
712 supported for compat reasons. Also, this directive now takes
713 an integer value between 1 and 63, or a boolean value. The
714 formerly supported '-1' value for disabling stays around for
717 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
718 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
719 RootDirectory= properties can now be set for transient
722 * The systemd-analyze tool gained a new "set-log-target" verb
723 to change the logging target the system manager logs to
724 dynamically during runtime. This is similar to how
725 "systemd-analyze set-log-level" already changes the log
728 * In nspawn /sys is now mounted as tmpfs, with only a selected
729 set of subdirectories mounted in from the real sysfs. This
730 enhances security slightly, and is useful for ensuring user
731 namespaces work correctly.
733 * Support for USB FunctionFS activation has been added. This
734 allows implementation of USB gadget services that are
735 activated as soon as they are requested, so that they don't
736 have to run continously, similar to classic socket
739 * The "systemctl exit" command now optionally takes an
740 additional parameter that sets the exit code to return from
741 the systemd manager when exiting. This is only relevant when
742 running the systemd user instance, or when running the
743 system instance in a container.
745 * sd-bus gained the new API calls sd_bus_path_encode_many()
746 and sd_bus_path_decode_many() that allow easy encoding and
747 decoding of multiple identifier strings inside a D-Bus
748 object path. Another new call sd_bus_default_flush_close()
749 has been added to flush and close per-thread default
752 * systemd-cgtop gained support for a -M/--machine= switch to
753 show the control groups within a certain container only.
755 * "systemctl kill" gained support for an optional --fail
756 switch. If specified the requested operation will fail of no
757 processes have been killed, because the unit had no
758 processes attached, or similar.
760 * A new systemd.crash_reboot=1 kernel command line option has
761 been added that triggers a reboot after crashing. This can
762 also be set through CrashReboot= in systemd.conf.
764 * The RuntimeDirectory= setting now understands unit
765 specifiers like %i or %f.
767 * A new (still internal) libary API sd-ipv4acd has been added,
768 that implements address conflict detection for IPv4. It's
769 based on code from sd-ipv4ll, and will be useful for
770 detecting DHCP address conflicts.
772 * File descriptors passed during socket activation may now be
773 named. A new API sd_listen_fds_with_names() is added to
774 access the names. The default names may be overridden,
775 either in the .socket file using the FileDescriptorName=
776 parameter, or by passing FDNAME= when storing the file
777 descriptors using sd_notify().
779 * systemd-networkd gained support for:
781 - Setting the IPv6 Router Advertisment settings via
782 IPv6AcceptRouterAdvertisements= in .network files.
784 - Configuring the HelloTimeSec=, MaxAgeSec= and
785 ForwardDelaySec= bridge parameters in .netdev files.
787 - Configuring PreferredSource= for static routes in
790 * The "ask-password" framework used to query for LUKS harddisk
791 passwords or SSL passwords during boot gained support for
792 caching passwords in the kernel keyring, if it is
793 available. This makes sure that the user only has to type in
794 a passphrase once if there are multiple objects to unlock
795 with the same one. Previously, such password caching was
796 available only when Plymouth was used; this moves the
797 caching logic into the systemd codebase itself. The
798 "systemd-ask-password" utility gained a new --keyname=
799 switch to control which kernel keyring key to use for
800 caching a password in. This functionality is also useful for
801 enabling display managers such as gdm to automatically
802 unlock the user's GNOME keyring if its passphrase, the
803 user's password and the harddisk password are the same, if
804 gdm-autologin is used.
806 * When downloading tar or raw images using "machinectl
807 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
808 file is now also downloaded, if it is available and stored
809 next to the image file.
811 * Units of type ".socket" gained a new boolean setting
812 Writable= which is only useful in conjunction with
813 ListenSpecial=. If true, enables opening the specified
814 special file in O_RDWR mode rather than O_RDONLY mode.
816 * systemd-rfkill has been reworked to become a singleton
817 service that is activated through /dev/rfkill on each rfkill
818 state change and saves the settings to disk. This way,
819 systemd-rfkill is now compatible with devices that exist
820 only intermittendly, and even restores state if the previous
821 system shutdown was abrupt rather than clean.
823 * The journal daemon gained support for vacuuming old journal
824 files controlled by the number of files that shall remain,
825 in addition to the already existing control by size and by
826 date. This is useful as journal interleaving performance
827 degrades with too many seperate journal files, and allows
828 putting an effective limit on them. The new setting defaults
829 to 100, but this may be changed by setting SystemMaxFiles=
830 and RuntimeMaxFiles= in journald.conf. Also, the
831 "journalctl" tool gained the new --vacuum-files= switch to
832 manually vacuum journal files to leave only the specified
833 number of files in place.
835 * udev will now create /dev/disk/by-path links for ATA devices
836 on kernels where that is supported.
838 * Galician, Serbian, Turkish and Korean translations were added.
840 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
841 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
842 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
843 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
844 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
845 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
846 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
847 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
848 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
849 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
850 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
851 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
852 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
853 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
854 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
855 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
856 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
857 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
863 * The DHCP implementation of systemd-networkd gained a set of
866 - The DHCP server now supports emitting DNS and NTP
867 information. It may be enabled and configured via
868 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
869 and NTP information is enabled, but no servers are
870 configured, the corresponding uplink information (if there
871 is any) is propagated.
873 - Server and client now support transmission and reception
874 of timezone information. It can be configured via the
875 newly introduced network options UseTimezone=,
876 EmitTimezone=, and Timezone=. Transmission of timezone
877 information is enabled between host and containers by
878 default now: the container will change its local timezone
879 to what the host has set.
881 - Lease timeouts can now be configured via
882 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
884 - The DHCP server improved on the stability of
885 leases. Clients are more likely to get the same lease
886 information back, even if the server loses state.
888 - The DHCP server supports two new configuration options to
889 control the lease address pool metrics, PoolOffset= and
892 * The encapsulation limit of tunnels in systemd-networkd may
893 now be configured via 'EncapsulationLimit='. It allows
894 modifying the maximum additional levels of encapsulation
895 that are permitted to be prepended to a packet.
897 * systemd now supports the concept of user buses replacing
898 session buses, if used with dbus-1.10 (and enabled via dbus
899 --enable-user-session). It previously only supported this on
900 kdbus-enabled systems, and this release expands this to
901 'dbus-daemon' systems.
903 * systemd-networkd now supports predictable interface names
906 * systemd now optionally supports the new Linux kernel
907 "unified" control group hierarchy. If enabled via the kernel
908 command-line option 'systemd.unified_cgroup_hierarchy=1',
909 systemd will try to mount the unified cgroup hierarchy
910 directly on /sys/fs/cgroup. If not enabled, or not
911 available, systemd will fall back to the legacy cgroup
912 hierarchy setup, as before. Host system and containers can
913 mix and match legacy and unified hierarchies as they
914 wish. nspawn understands the $UNIFIED_CROUP_HIERARCHY
915 environment variable to individually select the hierarchy to
916 use for executed containers. By default, nspawn will use the
917 unified hierarchy for the containers if the host uses the
918 unified hierarchy, and the legacy hierarchy otherwise.
919 Please note that at this point the unified hierarchy is an
920 experimental kernel feature and is likely to change in one
921 of the next kernel releases. Therefore, it should not be
922 enabled by default in downstream distributions yet. The
923 minimum required kernel version for the unified hierarchy to
924 work is 4.2. Note that when the unified hierarchy is used
925 for the first time delegated access to controllers is
926 safe. Because of this systemd-nspawn containers will get
927 access to controllers now, as will systemd user
928 sessions. This means containers and user sessions may now
929 manage their own resources, partitioning up what the system
932 * A new special scope unit "init.scope" has been introduced
933 that encapsulates PID 1 of the system. It may be used to
934 determine resource usage and enforce resource limits on PID
935 1 itself. PID 1 hence moved out of the root of the control
938 * The cgtop tool gained support for filtering out kernel
939 threads when counting tasks in a control group. Also, the
940 count of processes is now recursively summed up by
941 default. Two options -k and --recursive= have been added to
942 revert to old behaviour. The tool has also been updated to
943 work correctly in containers now.
945 * systemd-nspawn's --bind= and --bind-ro= options have been
946 extended to allow creation of non-recursive bind mounts.
948 * libsystemd gained two new calls sd_pid_get_cgroup() and
949 sd_peer_get_cgroup() which return the control group path of
950 a process or peer of a connected AF_UNIX socket. This
951 function call is particularly useful when implementing
952 delegated subtrees support in the control group hierarchy.
954 * The "sd-event" event loop API of libsystemd now supports
955 correct dequeuing of real-time signals, without losing
958 * When systemd requests a PolicyKit decision when managing
959 units it will now add additional fields to the request,
960 including unit name and desired operation. This enables more
961 powerful PolicyKit policies, that make decisions depending
964 * nspawn learnt support for .nspawn settings files, that may
965 accompany the image files or directories of containers, and
966 may contain additional settings for the container. This is
967 an alternative to configuring container parameters via the
970 Contributions from: Cristian Rodríguez, Daniel Mack, David
971 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
972 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
973 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
974 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
975 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
976 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
977 Andersen, Tom Gundersen, Torstein Husebø
983 * machinectl gained a new verb 'shell' which opens a fresh
984 shell on the target container or the host. It is similar to
985 the existing 'login' command of machinectl, but spawns the
986 shell directly without prompting for username or
987 password. The pseudo machine '.host' now refers to the local
988 host and is used by default. Hence, 'machinectl shell' can
989 be used as replacement for 'su -' which spawns a session as
990 a fresh systemd unit in a way that is fully isolated from
991 the originating session.
993 * systemd-networkd learned to cope with private-zone DHCP
994 options and allows other programs to query the values.
996 * SELinux access control when enabling/disabling units is no
997 longer enforced with this release. The previous
998 implementation was incorrect, and a new corrected
999 implementation is not yet available. As unit file operations
1000 are still protected via PolicyKit and D-Bus policy this is
1001 not a security problem. Yet, distributions which care about
1002 optimal SELinux support should probably not stabilize on
1005 * sd-bus gained support for matches of type "arg0has=", that
1006 test for membership of strings in string arrays sent in bus
1009 * systemd-resolved now dumps the contents of its DNS and LLMNR
1010 caches to the logs on reception of the SIGUSR1 signal. This
1011 is useful to debug DNS behaviour.
1013 * The coredumpctl tool gained a new --directory= option to
1014 operate on journal files in a specific directory.
1016 * "systemctl reboot" and related commands gained a new
1017 "--message=" option which may be used to set a free-text
1018 wall message when shutting down or rebooting the
1019 system. This message is also logged, which is useful for
1020 figuring out the reason for a reboot or shutdown a
1023 * The "systemd-resolve-host" tool's -i switch now takes
1024 network interface numbers as alternative to interface names.
1026 * A new unit file setting for services has been introduced:
1027 UtmpMode= allows configuration of how precisely systemd
1028 handles utmp and wtmp entries for the service if this is
1029 enabled. This allows writing services that appear similar to
1030 user sessions in the output of the "w", "who", "last" and
1033 * systemd-resolved will now locally synthesize DNS resource
1034 records for the "localhost" and "gateway" domains as well as
1035 the local hostname. This should ensure that clients querying
1036 RRs via resolved will get similar results as those going via
1037 NSS, if nss-myhostname is enabled.
1039 Contributions from: Alastair Hughes, Alex Crawford, Daniel
1040 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
1041 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
1042 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
1043 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
1044 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
1045 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
1046 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
1047 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
1048 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
1049 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
1050 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
1052 — Berlin, 2015-08-27
1056 * The systemd-efi-boot-generator functionality was merged into
1057 systemd-gpt-auto-generator.
1059 * systemd-networkd now supports Group Policy for vxlan
1060 devices. It can be enabled via the new boolean configuration
1061 option called 'GroupPolicyExtension='.
1063 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
1064 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
1065 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
1067 — Berlin, 2015-07-31
1071 * The python-systemd code has been removed from the systemd repository.
1072 A new repository has been created which accommodates the code from
1073 now on, and we kindly ask distributions to create a separate package
1074 for this: https://github.com/systemd/python-systemd
1076 * The systemd daemon will now reload its main configuration
1077 (/etc/systemd/system.conf) on daemon-reload.
1079 * sd-dhcp now exposes vendor specific extensions via
1080 sd_dhcp_lease_get_vendor_specific().
1082 * systemd-networkd gained a number of new configuration options.
1084 - A new boolean configuration option for TAP devices called
1085 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
1086 device, thus allowing to send and receive GSO packets.
1088 - A new tunnel configuration option called 'CopyDSCP='.
1089 If enabled, the DSCP field of ip6 tunnels is copied into the
1090 decapsulated packet.
1092 - A set of boolean bridge configuration options were added.
1093 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
1094 and 'UnicastFlood=' are now parsed by networkd and applied to the
1095 respective bridge link device via the respective IFLA_BRPORT_*
1098 - A new string configuration option to override the hostname sent
1099 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
1100 is true, networkd will use the configured hostname instead of the
1101 system hostname when sending DHCP requests.
1103 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
1104 networkd will configure the IPv6 flow-label of the tunnel device
1105 according to RFC2460.
1107 - The 'macvtap' virtual network devices are now supported, similar to
1108 the already supported 'macvlan' devices.
1110 * systemd-resolved now implements RFC5452 to improve resilience against
1111 cache poisoning. Additionally, source port randomization is enabled
1112 by default to further protect against DNS spoofing attacks.
1114 * nss-mymachines now supports translating UIDs and GIDs of running
1115 containers with user-namespaces enabled. If a container 'foo'
1116 translates a host uid 'UID' to the container uid 'TUID', then
1117 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
1118 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
1119 mapped as 'vg-foo-TGID'.
1121 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
1122 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
1123 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
1124 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
1125 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
1126 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
1127 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
1128 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
1129 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
1130 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
1132 — Berlin, 2015-07-29
1136 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
1137 There are no known issues with current sysfs, and udev does not need
1138 or should be used to work around such bugs.
1140 * udev does no longer enable USB HID power management. Several reports
1141 indicate, that some devices cannot handle that setting.
1143 * The udev accelerometer helper was removed. The functionality
1144 is now fully included in iio-sensor-proxy. But this means,
1145 older iio-sensor-proxy versions will no longer provide
1146 accelerometer/orientation data with this systemd version.
1147 Please upgrade iio-sensor-proxy to version 1.0.
1149 * networkd gained a new configuration option IPv6PrivacyExtensions=
1150 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
1151 for Stateless Address") on selected networks.
1153 * For the sake of fewer build-time dependencies and less code in the
1154 main repository, the python bindings are about to be removed in the
1155 next release. A new repository has been created which accommodates
1156 the code from now on, and we kindly ask distributions to create a
1157 separate package for this. The removal will take place in v223.
1159 https://github.com/systemd/python-systemd
1161 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
1162 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
1163 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
1164 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
1165 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
1166 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
1167 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
1168 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
1169 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
1170 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
1172 — Berlin, 2015-07-07
1176 * The sd-bus.h and sd-event.h APIs have now been declared
1177 stable and have been added to the official interface of
1178 libsystemd.so. sd-bus implements an alternative D-Bus client
1179 library, that is relatively easy to use, very efficient and
1180 supports both classic D-Bus as well as kdbus as transport
1181 backend. sd-event is a generic event loop abstraction that
1182 is built around Linux epoll, but adds features such as event
1183 prioritization or efficient timer handling. Both APIs are good
1184 choices for C programs looking for a bus and/or event loop
1185 implementation that is minimal and does not have to be
1186 portable to other kernels.
1188 * kdbus support is no longer compile-time optional. It is now
1189 always built-in. However, it can still be disabled at
1190 runtime using the kdbus=0 kernel command line setting, and
1191 that setting may be changed to default to off, by specifying
1192 --disable-kdbus at build-time. Note though that the kernel
1193 command line setting has no effect if the kdbus.ko kernel
1194 module is not installed, in which case kdbus is (obviously)
1195 also disabled. We encourage all downstream distributions to
1196 begin testing kdbus by adding it to the kernel images in the
1197 development distributions, and leaving kdbus support in
1200 * The minimal required util-linux version has been bumped to
1203 * Support for chkconfig (--enable-chkconfig) was removed in
1204 favor of calling an abstraction tool
1205 /lib/systemd/systemd-sysv-install. This needs to be
1206 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
1207 in README for details.
1209 * If there's a systemd unit and a SysV init script for the
1210 same service name, and the user executes "systemctl enable"
1211 for it (or a related call), then this will now enable both
1212 (or execute the related operation on both), not just the
1215 * The libudev API documentation has been converted from gtkdoc
1218 * gudev has been removed from the systemd tree, it is now an
1221 * The systemd-cgtop tool learnt a new --raw switch to generate
1222 "raw" (machine parsable) output.
1224 * networkd's IPForwarding= .network file setting learnt the
1225 new setting "kernel", which ensures that networkd does not
1226 change the IP forwarding sysctl from the default kernel
1229 * The systemd-logind bus API now exposes a new boolean
1230 property "Docked" that reports whether logind considers the
1231 system "docked", i.e. connected to a docking station or not.
1233 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
1234 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
1235 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
1236 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
1237 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
1238 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
1239 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
1240 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
1241 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
1242 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
1243 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
1244 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
1245 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
1246 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
1247 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
1248 Fink, Zbigniew Jędrzejewski-Szmek
1250 — Berlin, 2015-06-19
1254 * The gudev library has been extracted into a separate repository
1255 available at: https://git.gnome.org/browse/libgudev/
1256 It is now managed as part of the Gnome project. Distributions
1257 are recommended to pass --disable-gudev to systemd and use
1258 gudev from the Gnome project instead. gudev is still included
1259 in systemd, for now. It will be removed soon, though. Please
1260 also see the announcement-thread on systemd-devel:
1261 http://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
1263 * systemd now exposes a CPUUsageNSec= property for each
1264 service unit on the bus, that contains the overall consumed
1265 CPU time of a service (the sum of what each process of the
1266 service consumed). This value is only available if
1267 CPUAccounting= is turned on for a service, and is then shown
1268 in the "systemctl status" output.
1270 * Support for configuring alternative mappings of the old SysV
1271 runlevels to systemd targets has been removed. They are now
1272 hardcoded in a way that runlevels 2, 3, 4 all map to
1273 multi-user.target and 5 to graphical.target (which
1274 previously was already the default behaviour).
1276 * The auto-mounter logic gained support for mount point
1277 expiry, using a new TimeoutIdleSec= setting in .automount
1278 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
1280 * The EFI System Partition (ESP) as mounted to /boot by
1281 systemd-efi-boot-generator will now be unmounted
1282 automatically after 2 minutes of not being used. This should
1283 minimize the risk of ESP corruptions.
1285 * New /etc/fstab options x-systemd.requires= and
1286 x-systemd.requires-mounts-for= are now supported to express
1287 additional dependencies for mounts. This is useful for
1288 journalling file systems that support external journal
1289 devices or overlay file systems that require underlying file
1290 systems to be mounted.
1292 * systemd does not support direct live-upgrades (via systemctl
1293 daemon-reexec) from versions older than v44 anymore. As no
1294 distribution we are aware of shipped such old versions in a
1295 stable release this should not be problematic.
1297 * When systemd forks off a new per-connection service instance
1298 it will now set the $REMOTE_ADDR environment variable to the
1299 remote IP address, and $REMOTE_PORT environment variable to
1300 the remote IP port. This behaviour is similar to the
1301 corresponding environment variables defined by CGI.
1303 * systemd-networkd gained support for uplink failure
1304 detection. The BindCarrier= option allows binding interface
1305 configuration dynamically to the link sense of other
1306 interfaces. This is useful to achieve behaviour like in
1309 * systemd-networkd gained support for configuring the DHCP
1310 client identifier to use when requesting leases.
1312 * systemd-networkd now has a per-network UseNTP= option to
1313 configure whether NTP server information acquired via DHCP
1314 is passed on to services like systemd-timesyncd.
1316 * systemd-networkd gained support for vti6 tunnels.
1318 * Note that systemd-networkd manages the sysctl variable
1319 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
1320 it is configured for since v219. The variable controls IP
1321 forwarding, and is a per-interface alternative to the global
1322 /proc/sys/net/ipv[46]/ip_forward. This setting is
1323 configurable in the IPForward= option, which defaults to
1324 "no". This means if networkd is used for an interface it is
1325 no longer sufficient to set the global sysctl option to turn
1326 on IP forwarding! Instead, the .network file option
1327 IPForward= needs to be turned on! Note that the
1328 implementation of this behaviour was broken in v219 and has
1331 * Many bonding and vxlan options are now configurable in
1334 * systemd-nspawn gained a new --property= setting to set unit
1335 properties for the container scope. This is useful for
1336 setting resource parameters (e.g "CPUShares=500") on
1337 containers started from the command line.
1339 * systemd-nspawn gained a new --private-users= switch to make
1340 use of user namespacing available on recent Linux kernels.
1342 * systemd-nspawn may now be called as part of a shell pipeline
1343 in which case the pipes used for stdin and stdout are passed
1344 directly to the process invoked in the container, without
1345 indirection via a pseudo tty.
1347 * systemd-nspawn gained a new switch to control the UNIX
1348 signal to use when killing the init process of the container
1351 * systemd-nspawn gained a new --overlay= switch for mounting
1352 overlay file systems into the container using the new kernel
1355 * When a container image is imported via systemd-importd and
1356 the host file system is not btrfs, a loopback block device
1357 file is created in /var/lib/machines.raw with a btrfs file
1358 system inside. It is then mounted to /var/lib/machines to
1359 enable btrfs features for container management. The loopback
1360 file and btrfs file system is grown as needed when container
1361 images are imported via systemd-importd.
1363 * systemd-machined/systemd-importd gained support for btrfs
1364 quota, to enforce container disk space limits on disk. This
1365 is exposed in "machinectl set-limit".
1367 * systemd-importd now can import containers from local .tar,
1368 .raw and .qcow2 images, and export them to .tar and .raw. It
1369 can also import dkr v2 images now from the network (on top
1372 * systemd-importd gained support for verifying downloaded
1373 images with gpg2 (previously only gpg1 was supported).
1375 * systemd-machined, systemd-logind, systemd: most bus calls
1376 are now accessible to unprivileged processes via
1377 PolicyKit. Also, systemd-logind will now allow users to kill
1378 their own sessions without further privileges or
1381 * systemd-shutdownd has been removed. This service was
1382 previously responsible for implementing scheduled shutdowns
1383 as exposed in /usr/bin/shutdown's time parameter. This
1384 functionality has now been moved into systemd-logind and is
1385 accessible via a bus interface.
1387 * "systemctl reboot" gained a new switch --firmware-setup that
1388 can be used to reboot into the EFI firmware setup, if that
1389 is available. systemd-logind now exposes an API on the bus
1390 to trigger such reboots, in case graphical desktop UIs want
1391 to cover this functionality.
1393 * "systemctl enable", "systemctl disable" and "systemctl mask"
1394 now support a new "--now" switch. If specified the units
1395 that are enabled will also be started, and the ones
1396 disabled/masked also stopped.
1398 * The Gummiboot EFI boot loader tool has been merged into
1399 systemd, and renamed to "systemd-boot". The bootctl tool has been
1400 updated to support systemd-boot.
1402 * An EFI kernel stub has been added that may be used to create
1403 kernel EFI binaries that contain not only the actual kernel,
1404 but also an initrd, boot splash, command line and OS release
1405 information. This combined binary can then be signed as a
1406 single image, so that the firmware can verify it all in one
1407 step. systemd-boot has special support for EFI binaries created
1408 like this and can extract OS release information from them
1409 and show them in the boot menu. This functionality is useful
1410 to implement cryptographically verified boot schemes.
1412 * Optional support has been added to systemd-fsck to pass
1413 fsck's progress report to an AF_UNIX socket in the file
1416 * udev will no longer create device symlinks for all block
1417 devices by default. A blacklist for excluding special block
1418 devices from this logic has been turned into a whitelist
1419 that requires picking block devices explicitly that require
1422 * A new (currently still internal) API sd-device.h has been
1423 added to libsystemd. This modernized API is supposed to
1424 replace libudev eventually. In fact, already much of libudev
1425 is now just a wrapper around sd-device.h.
1427 * A new hwdb database for storing metadata about pointing
1428 stick devices has been added.
1430 * systemd-tmpfiles gained support for setting file attributes
1431 similar to the "chattr" tool with new 'h' and 'H' lines.
1433 * systemd-journald will no longer unconditionally set the
1434 btrfs NOCOW flag on new journal files. This is instead done
1435 with tmpfiles snippet using the new 'h' line type. This
1436 allows easy disabling of this logic, by masking the
1437 journal-nocow.conf tmpfiles file.
1439 * systemd-journald will now translate audit message types to
1440 human readable identifiers when writing them to the
1441 journal. This should improve readability of audit messages.
1443 * The LUKS logic gained support for the offset= and skip=
1444 options in /etc/crypttab, as previously implemented by
1447 * /usr/lib/os-release gained a new optional field VARIANT= for
1448 distributions that support multiple variants (such as a
1449 desktop edition, a server edition, ...)
1451 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
1452 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
1453 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
1454 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
1455 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
1456 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
1457 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
1458 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
1459 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
1460 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
1461 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
1462 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
1463 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
1464 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
1465 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
1466 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
1467 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
1468 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
1469 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
1470 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
1471 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
1472 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
1473 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
1474 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
1475 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
1476 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
1477 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
1479 — Berlin, 2015-05-22
1483 * Introduce a new API "sd-hwdb.h" for querying the hardware
1484 metadata database. With this minimal interface one can query
1485 and enumerate the udev hwdb, decoupled from the old libudev
1486 library. libudev's interface for this is now only a wrapper
1487 around sd-hwdb. A new tool systemd-hwdb has been added to
1488 interface with and update the database.
1490 * When any of systemd's tools copies files (for example due to
1491 tmpfiles' C lines) a btrfs reflink will attempted first,
1492 before bytewise copying is done.
1494 * systemd-nspawn gained a new --ephemeral switch. When
1495 specified a btrfs snapshot is taken of the container's root
1496 directory, and immediately removed when the container
1497 terminates again. Thus, a container can be started whose
1498 changes never alter the container's root directory, and are
1499 lost on container termination. This switch can also be used
1500 for starting a container off the root file system of the
1501 host without affecting the host OS. This switch is only
1502 available on btrfs file systems.
1504 * systemd-nspawn gained a new --template= switch. It takes the
1505 path to a container tree to use as template for the tree
1506 specified via --directory=, should that directory be
1507 missing. This allows instantiating containers dynamically,
1508 on first run. This switch is only available on btrfs file
1511 * When a .mount unit refers to a mount point on which multiple
1512 mounts are stacked, and the .mount unit is stopped all of
1513 the stacked mount points will now be unmounted until no
1514 mount point remains.
1516 * systemd now has an explicit notion of supported and
1517 unsupported unit types. Jobs enqueued for unsupported unit
1518 types will now fail with an "unsupported" error code. More
1519 specifically .swap, .automount and .device units are not
1520 supported in containers, .busname units are not supported on
1521 non-kdbus systems. .swap and .automount are also not
1522 supported if their respective kernel compile time options
1525 * machinectl gained support for two new "copy-from" and
1526 "copy-to" commands for copying files from a running
1527 container to the host or vice versa.
1529 * machinectl gained support for a new "bind" command to bind
1530 mount host directories into local containers. This is
1531 currently only supported for nspawn containers.
1533 * networkd gained support for configuring bridge forwarding
1534 database entries (fdb) from .network files.
1536 * A new tiny daemon "systemd-importd" has been added that can
1537 download container images in tar, raw, qcow2 or dkr formats,
1538 and make them available locally in /var/lib/machines, so
1539 that they can run as nspawn containers. The daemon can GPG
1540 verify the downloads (not supported for dkr, since it has no
1541 provisions for verifying downloads). It will transparently
1542 decompress bz2, xz, gzip compressed downloads if necessary,
1543 and restore sparse files on disk. The daemon uses privilege
1544 separation to ensure the actual download logic runs with
1545 fewer privileges than the daemon itself. machinectl has
1546 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
1547 make the functionality of importd available to the
1548 user. With this in place the Fedora and Ubuntu "Cloud"
1549 images can be downloaded and booted as containers unmodified
1550 (the Fedora images lack the appropriate GPG signature files
1551 currently, so they cannot be verified, but this will change
1552 soon, hopefully). Note that downloading images is currently
1553 only fully supported on btrfs.
1555 * machinectl is now able to list container images found in
1556 /var/lib/machines, along with some metadata about sizes of
1557 disk and similar. If the directory is located on btrfs and
1558 quota is enabled, this includes quota display. A new command
1559 "image-status" has been added that shows additional
1560 information about images.
1562 * machinectl is now able to clone container images
1563 efficiently, if the underlying file system (btrfs) supports
1564 it, with the new "machinectl clone" command. It also
1565 gained commands for renaming and removing images, as well as
1566 marking them read-only or read-write (supported also on
1567 legacy file systems).
1569 * networkd gained support for collecting LLDP network
1570 announcements, from hardware that supports this. This is
1571 shown in networkctl output.
1573 * systemd-run gained support for a new -t (--pty) switch for
1574 invoking a binary on a pty whose input and output is
1575 connected to the invoking terminal. This allows executing
1576 processes as system services while interactively
1577 communicating with them via the terminal. Most interestingly
1578 this is supported across container boundaries. Invoking
1579 "systemd-run -t /bin/bash" is an alternative to running a
1580 full login session, the difference being that the former
1581 will not register a session, nor go through the PAM session
1584 * tmpfiles gained support for a new "v" line type for creating
1585 btrfs subvolumes. If the underlying file system is a legacy
1586 file system, this automatically degrades to creating a
1587 normal directory. Among others /var/lib/machines is now
1588 created like this at boot, should it be missing.
1590 * The directory /var/lib/containers/ has been deprecated and
1591 been replaced by /var/lib/machines. The term "machines" has
1592 been used in the systemd context as generic term for both
1593 VMs and containers, and hence appears more appropriate for
1594 this, as the directory can also contain raw images bootable
1597 * systemd-nspawn when invoked with -M but without --directory=
1598 or --image= is now capable of searching for the container
1599 root directory, subvolume or disk image automatically, in
1600 /var/lib/machines. systemd-nspawn@.service has been updated
1601 to make use of this, thus allowing it to be used for raw
1604 * A new machines.target unit has been introduced that is
1605 supposed to group all containers/VMs invoked as services on
1606 the system. systemd-nspawn@.service has been updated to
1607 integrate with that.
1609 * machinectl gained a new "start" command, for invoking a
1610 container as a service. "machinectl start foo" is mostly
1611 equivalent to "systemctl start systemd-nspawn@foo.service",
1612 but handles escaping in a nicer way.
1614 * systemd-nspawn will now mount most of the cgroupfs tree
1615 read-only into each container, with the exception of the
1616 container's own subtree in the name=systemd hierarchy.
1618 * journald now sets the special FS_NOCOW file flag for its
1619 journal files. This should improve performance on btrfs, by
1620 avoiding heavy fragmentation when journald's write-pattern
1621 is used on COW file systems. It degrades btrfs' data
1622 integrity guarantees for the files to the same levels as for
1623 ext3/ext4 however. This should be OK though as journald does
1624 its own data integrity checks and all its objects are
1625 checksummed on disk. Also, journald should handle btrfs disk
1626 full events a lot more gracefully now, by processing SIGBUS
1627 errors, and not relying on fallocate() anymore.
1629 * When journald detects that journal files it is writing to
1630 have been deleted it will immediately start new journal
1633 * systemd now provides a way to store file descriptors
1634 per-service in PID 1.This is useful for daemons to ensure
1635 that fds they require are not lost during a daemon
1636 restart. The fds are passed to the daemon on the next
1637 invocation in the same way socket activation fds are
1638 passed. This is now used by journald to ensure that the
1639 various sockets connected to all the system's stdout/stderr
1640 are not lost when journald is restarted. File descriptors
1641 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
1642 an extension to sd_notify(). Note that a limit is enforced
1643 on the number of fds a service can store in PID 1, and it
1644 defaults to 0, so that no fds may be stored, unless this is
1645 explicitly turned on.
1647 * The default TERM variable to use for units connected to a
1648 terminal, when no other value is explicitly is set is now
1649 vt220 rather than vt102. This should be fairly safe still,
1650 but allows PgUp/PgDn work.
1652 * The /etc/crypttab option header= as known from Debian is now
1655 * "loginctl user-status" and "loginctl session-status" will
1656 now show the last 10 lines of log messages of the
1657 user/session following the status output. Similar,
1658 "machinectl status" will show the last 10 log lines
1659 associated with a virtual machine or container
1660 service. (Note that this is usually not the log messages
1661 done in the VM/container itself, but simply what the
1662 container manager logs. For nspawn this includes all console
1665 * "loginctl session-status" without further argument will now
1666 show the status of the session of the caller. Similar,
1667 "lock-session", "unlock-session", "activate",
1668 "enable-linger", "disable-linger" may now be called without
1669 session/user parameter in which case they apply to the
1670 caller's session/user.
1672 * An X11 session scriptlet is now shipped that uploads
1673 $DISPLAY and $XAUTHORITY into the environment of the systemd
1674 --user daemon if a session begins. This should improve
1675 compatibility with X11 enabled applications run as systemd
1678 * Generators are now subject to masking via /etc and /run, the
1679 same way as unit files.
1681 * networkd .network files gained support for configuring
1682 per-link IPv4/IPv6 packet forwarding as well as IPv4
1683 masquerading. This is by default turned on for veth links to
1684 containers, as registered by systemd-nspawn. This means that
1685 nspawn containers run with --network-veth will now get
1686 automatic routed access to the host's networks without any
1687 further configuration or setup, as long as networkd runs on
1690 * systemd-nspawn gained the --port= (-p) switch to expose TCP
1691 or UDP posts of a container on the host. With this in place
1692 it is possible to run containers with private veth links
1693 (--network-veth), and have their functionality exposed on
1694 the host as if their services were running directly on the
1697 * systemd-nspawn's --network-veth switch now gained a short
1698 version "-n", since with the changes above it is now truly
1699 useful out-of-the-box. The systemd-nspawn@.service has been
1700 updated to make use of it too by default.
1702 * systemd-nspawn will now maintain a per-image R/W lock, to
1703 ensure that the same image is not started more than once
1704 writable. (It's OK to run an image multiple times
1705 simultaneously in read-only mode.)
1707 * systemd-nspawn's --image= option is now capable of
1708 dissecting and booting MBR and GPT disk images that contain
1709 only a single active Linux partition. Previously it
1710 supported only GPT disk images with proper GPT type
1711 IDs. This allows running cloud images from major
1712 distributions directly with systemd-nspawn, without
1715 * In addition to collecting mouse dpi data in the udev
1716 hardware database, there's now support for collecting angle
1717 information for mouse scroll wheels. The database is
1718 supposed to guarantee similar scrolling behavior on mice
1719 that it knows about. There's also support for collecting
1720 information about Touchpad types.
1722 * udev's input_id built-in will now also collect touch screen
1723 dimension data and attach it to probed devices.
1725 * /etc/os-release gained support for a Distribution Privacy
1728 * networkd gained support for creating "ipvlan", "gretap",
1729 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
1731 * systemd-tmpfiles gained support for "a" lines for setting
1734 * systemd-nspawn will now mount /tmp in the container to
1735 tmpfs, automatically.
1737 * systemd now exposes the memory.usage_in_bytes cgroup
1738 attribute and shows it for each service in the "systemctl
1739 status" output, if available.
1741 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
1742 immediate reboot is triggered. This useful if shutdown is
1743 hung and is unable to complete, to expedite the
1744 operation. Note that this kind of reboot will still unmount
1745 all file systems, and hence should not result in fsck being
1748 * A .device unit for an optical block device will now be
1749 considered active only when a medium is in the drive. Also,
1750 mount units are now bound to their backing devices thus
1751 triggering automatic unmounting when devices become
1752 unavailable. With this in place systemd will now
1753 automatically unmount left-over mounts when a CD-ROM is
1754 ejected or an USB stick is yanked from the system.
1756 * networkd-wait-online now has support for waiting for
1757 specific interfaces only (with globbing), and for giving up
1758 after a configurable timeout.
1760 * networkd now exits when idle. It will be automatically
1761 restarted as soon as interfaces show up, are removed or
1762 change state. networkd will stay around as long as there is
1763 at least one DHCP state machine or similar around, that keep
1766 * networkd may now configure IPv6 link-local addressing in
1767 addition to IPv4 link-local addressing.
1769 * The IPv6 "token" for use in SLAAC may now be configured for
1770 each .network interface in networkd.
1772 * Routes configured with networkd may now be assigned a scope
1775 * networkd's [Match] sections now support globbing and lists
1776 of multiple space-separated matches per item.
1778 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
1779 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
1780 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
1781 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
1782 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
1783 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
1784 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
1785 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
1786 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
1787 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
1788 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
1789 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
1790 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
1791 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
1792 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
1793 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
1794 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
1795 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
1796 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
1797 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
1798 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
1799 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
1800 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
1801 Hoffmann, Zbigniew Jędrzejewski-Szmek
1803 — Berlin, 2015-02-16
1807 * When querying unit file enablement status (for example via
1808 "systemctl is-enabled"), a new state "indirect" is now known
1809 which indicates that a unit might not be enabled itself, but
1810 another unit listed in its Also= setting might be.
1812 * Similar to the various existing ConditionXYZ= settings for
1813 units, there are now matching AssertXYZ= settings. While
1814 failing conditions cause a unit to be skipped, but its job
1815 to succeed, failing assertions declared like this will cause
1816 a unit start operation and its job to fail.
1818 * hostnamed now knows a new chassis type "embedded".
1820 * systemctl gained a new "edit" command. When used on a unit
1821 file, this allows extending unit files with .d/ drop-in
1822 configuration snippets or editing the full file (after
1823 copying it from /usr/lib to /etc). This will invoke the
1824 user's editor (as configured with $EDITOR), and reload the
1825 modified configuration after editing.
1827 * "systemctl status" now shows the suggested enablement state
1828 for a unit, as declared in the (usually vendor-supplied)
1829 system preset files.
1831 * nss-myhostname will now resolve the single-label host name
1832 "gateway" to the locally configured default IP routing
1833 gateways, ordered by their metrics. This assigns a stable
1834 name to the used gateways, regardless which ones are
1835 currently configured. Note that the name will only be
1836 resolved after all other name sources (if nss-myhostname is
1837 configured properly) and should hence not negatively impact
1838 systems that use the single-label host name "gateway" in
1841 * systemd-inhibit now allows filtering by mode when listing
1844 * Scope and service units gained a new "Delegate" boolean
1845 property, which, when set, allows processes running inside the
1846 unit to further partition resources. This is primarily
1847 useful for systemd user instances as well as container
1850 * journald will now pick up audit messages directly from
1851 the kernel, and log them like any other log message. The
1852 audit fields are split up and fully indexed. This means that
1853 journalctl in many ways is now a (nicer!) alternative to
1854 ausearch, the traditional audit client. Note that this
1855 implements only a minimal audit client. If you want the
1856 special audit modes like reboot-on-log-overflow, please use
1857 the traditional auditd instead, which can be used in
1858 parallel to journald.
1860 * The ConditionSecurity= unit file option now understands the
1861 special string "audit" to check whether auditing is
1864 * journalctl gained two new commands --vacuum-size= and
1865 --vacuum-time= to delete old journal files until the
1866 remaining ones take up no more than the specified size on disk,
1867 or are not older than the specified time.
1869 * A new, native PPPoE library has been added to sd-network,
1870 systemd's library of light-weight networking protocols. This
1871 library will be used in a future version of networkd to
1872 enable PPPoE communication without an external pppd daemon.
1874 * The busctl tool now understands a new "capture" verb that
1875 works similar to "monitor", but writes a packet capture
1876 trace to STDOUT that can be redirected to a file which is
1877 compatible with libcap's capture file format. This can then
1878 be loaded in Wireshark and similar tools to inspect bus
1881 * The busctl tool now understands a new "tree" verb that shows
1882 the object trees of a specific service on the bus, or of all
1885 * The busctl tool now understands a new "introspect" verb that
1886 shows all interfaces and members of objects on the bus,
1887 including their signature and values. This is particularly
1888 useful to get more information about bus objects shown by
1889 the new "busctl tree" command.
1891 * The busctl tool now understands new verbs "call",
1892 "set-property" and "get-property" for invoking bus method
1893 calls, setting and getting bus object properties in a
1896 * busctl gained a new --augment-creds= argument that controls
1897 whether the tool shall augment credential information it
1898 gets from the bus with data from /proc, in a possibly
1901 * nspawn's --link-journal= switch gained two new values
1902 "try-guest" and "try-host" that work like "guest" and
1903 "host", but do not fail if the host has no persistent
1904 journalling enabled. -j is now equivalent to
1905 --link-journal=try-guest.
1907 * macvlan network devices created by nspawn will now have
1908 stable MAC addresses.
1910 * A new SmackProcessLabel= unit setting has been added, which
1911 controls the SMACK security label processes forked off by
1912 the respective unit shall use.
1914 * If compiled with --enable-xkbcommon, systemd-localed will
1915 verify x11 keymap settings by compiling the given keymap. It
1916 will spew out warnings if the compilation fails. This
1917 requires libxkbcommon to be installed.
1919 * When a coredump is collected, a larger number of metadata
1920 fields is now collected and included in the journal records
1921 created for it. More specifically, control group membership,
1922 environment variables, memory maps, working directory,
1923 chroot directory, /proc/$PID/status, and a list of open file
1924 descriptors is now stored in the log entry.
1926 * The udev hwdb now contains DPI information for mice. For
1929 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
1931 * All systemd programs that read standalone configuration
1932 files in /etc now also support a corresponding series of
1933 .conf.d configuration directories in /etc/, /run/,
1934 /usr/local/lib/, /usr/lib/, and (if configured with
1935 --enable-split-usr) /lib/. In particular, the following
1936 configuration files now have corresponding configuration
1937 directories: system.conf user.conf, logind.conf,
1938 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
1939 resolved.conf, timesyncd.conf, journal-remote.conf, and
1940 journal-upload.conf. Note that distributions should use the
1941 configuration directories in /usr/lib/; the directories in
1942 /etc/ are reserved for the system administrator.
1944 * systemd-rfkill will no longer take the rfkill device name
1945 into account when storing rfkill state on disk, as the name
1946 might be dynamically assigned and not stable. Instead, the
1947 ID_PATH udev variable combined with the rfkill type (wlan,
1948 bluetooth, ...) is used.
1950 * A new service systemd-machine-id-commit.service has been
1951 added. When used on systems where /etc is read-only during
1952 boot, and /etc/machine-id is not initialized (but an empty
1953 file), this service will copy the temporary machine ID
1954 created as replacement into /etc after the system is fully
1955 booted up. This is useful for systems that are freshly
1956 installed with a non-initialized machine ID, but should get
1957 a fixed machine ID for subsequent boots.
1959 * networkd's .netdev files now provide a large set of
1960 configuration parameters for VXLAN devices. Similarly, the
1961 bridge port cost parameter is now configurable in .network
1962 files. There's also new support for configuring IP source
1963 routing. networkd .link files gained support for a new
1964 OriginalName= match that is useful to match against the
1965 original interface name the kernel assigned. .network files
1966 may include MTU= and MACAddress= fields for altering the MTU
1967 and MAC address while being connected to a specific network
1970 * The LUKS logic gained supported for configuring
1971 UUID-specific key files. There's also new support for naming
1972 LUKS device from the kernel command line, using the new
1973 luks.name= argument.
1975 * Timer units may now be transiently created via the bus API
1976 (this was previously already available for scope and service
1977 units). In addition it is now possible to create multiple
1978 transient units at the same time with a single bus call. The
1979 "systemd-run" tool has been updated to make use of this for
1980 running commands on a specified time, in at(1)-style.
1982 * tmpfiles gained support for "t" lines, for assigning
1983 extended attributes to files. Among other uses this may be
1984 used to assign SMACK labels to files.
1986 Contributions from: Alin Rauta, Alison Chaiken, Andrej
1987 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
1988 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
1989 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
1990 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
1991 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
1992 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
1993 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
1994 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
1995 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
1996 Michael Biebl, Michael Chapman, Michael Marineau, Michal
1997 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
1998 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
1999 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
2000 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
2001 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2002 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
2003 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
2005 — Berlin, 2014-12-10
2009 * journalctl gained the new options -t/--identifier= to match
2010 on the syslog identifier (aka "tag"), as well as --utc to
2011 show log timestamps in the UTC timezone. journalctl now also
2012 accepts -n/--lines=all to disable line capping in a pager.
2014 * journalctl gained a new switch, --flush, that synchronously
2015 flushes logs from /run/log/journal to /var/log/journal if
2016 persistent storage is enabled. systemd-journal-flush.service
2017 now waits until the operation is complete.
2019 * Services can notify the manager before they start a reload
2020 (by sending RELOADING=1) or shutdown (by sending
2021 STOPPING=1). This allows the manager to track and show the
2022 internal state of daemons and closes a race condition when
2023 the process is still running but has closed its D-Bus
2026 * Services with Type=oneshot do not have to have any ExecStart
2029 * User units are now loaded also from
2030 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
2031 /run/systemd/user directory that was already previously
2032 supported, but is under the control of the user.
2034 * Job timeouts (i.e. time-outs on the time a job that is
2035 queued stays in the run queue) can now optionally result in
2036 immediate reboot or power-off actions (JobTimeoutAction= and
2037 JobTimeoutRebootArgument=). This is useful on ".target"
2038 units, to limit the maximum time a target remains
2039 undispatched in the run queue, and to trigger an emergency
2040 operation in such a case. This is now used by default to
2041 turn off the system if boot-up (as defined by everything in
2042 basic.target) hangs and does not complete for at least
2043 15min. Also, if power-off or reboot hang for at least 30min
2044 an immediate power-off/reboot operation is triggered. This
2045 functionality is particularly useful to increase reliability
2046 on embedded devices, but also on laptops which might
2047 accidentally get powered on when carried in a backpack and
2048 whose boot stays stuck in a hard disk encryption passphrase
2051 * systemd-logind can be configured to also handle lid switch
2052 events even when the machine is docked or multiple displays
2053 are attached (HandleLidSwitchDocked= option).
2055 * A helper binary and a service have been added which can be
2056 used to resume from hibernation in the initramfs. A
2057 generator will parse the resume= option on the kernel
2058 command line to trigger resume.
2060 * A user console daemon systemd-consoled has been
2061 added. Currently, it is a preview, and will so far open a
2062 single terminal on each session of the user marked as
2063 Desktop=systemd-console.
2065 * Route metrics can be specified for DHCP routes added by
2068 * The SELinux context of socket-activated services can be set
2069 from the information provided by the networking stack
2070 (SELinuxContextFromNet= option).
2072 * Userspace firmware loading support has been removed and
2073 the minimum supported kernel version is thus bumped to 3.7.
2075 * Timeout for udev workers has been increased from 1 to 3
2076 minutes, but a warning will be printed after 1 minute to
2077 help diagnose kernel modules that take a long time to load.
2079 * Udev rules can now remove tags on devices with TAG-="foobar".
2081 * systemd's readahead implementation has been removed. In many
2082 circumstances it didn't give expected benefits even for
2083 rotational disk drives and was becoming less relevant in the
2084 age of SSDs. As none of the developers has been using
2085 rotating media anymore, and nobody stepped up to actively
2086 maintain this component of systemd it has now been removed.
2088 * Swap units can use Options= to specify discard options.
2089 Discard options specified for swaps in /etc/fstab are now
2092 * Docker containers are now detected as a separate type of
2095 * The Password Agent protocol gained support for queries where
2096 the user input is shown, useful e.g. for user names.
2097 systemd-ask-password gained a new --echo option to turn that
2100 * The default sysctl.d/ snippets will now set:
2102 net.core.default_qdisc = fq_codel
2104 This selects Fair Queuing Controlled Delay as the default
2105 queuing discipline for network interfaces. fq_codel helps
2106 fight the network bufferbloat problem. It is believed to be
2107 a good default with no tuning required for most workloads.
2108 Downstream distributions may override this choice. On 10Gbit
2109 servers that do not do forwarding, "fq" may perform better.
2110 Systems without a good clocksource should use "pfifo_fast".
2112 * If kdbus is enabled during build a new option BusPolicy= is
2113 available for service units, that allows locking all service
2114 processes into a stricter bus policy, in order to limit
2115 access to various bus services, or even hide most of them
2116 from the service's view entirely.
2118 * networkctl will now show the .network and .link file
2119 networkd has applied to a specific interface.
2121 * sd-login gained a new API call sd_session_get_desktop() to
2122 query which desktop environment has been selected for a
2125 * UNIX utmp support is now compile-time optional to support
2126 legacy-free systems.
2128 * systemctl gained two new commands "add-wants" and
2129 "add-requires" for pulling in units from specific targets
2132 * If the word "rescue" is specified on the kernel command line
2133 the system will now boot into rescue mode (aka
2134 rescue.target), which was previously available only by
2135 specifying "1" or "systemd.unit=rescue.target" on the kernel
2136 command line. This new kernel command line option nicely
2137 mirrors the already existing "emergency" kernel command line
2140 * New kernel command line options mount.usr=, mount.usrflags=,
2141 mount.usrfstype= have been added that match root=, rootflags=,
2142 rootfstype= but allow mounting a specific file system to
2145 * The $NOTIFY_SOCKET is now also passed to control processes of
2146 services, not only the main process.
2148 * This version reenables support for fsck's -l switch. This
2149 means at least version v2.25 of util-linux is required for
2150 operation, otherwise dead-locks on device nodes may
2151 occur. Again: you need to update util-linux to at least
2152 v2.25 when updating systemd to v217.
2154 * The "multi-seat-x" tool has been removed from systemd, as
2155 its functionality has been integrated into X servers 1.16,
2156 and the tool is hence redundant. It is recommended to update
2157 display managers invoking this tool to simply invoke X
2158 directly from now on, again.
2160 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
2161 message flag has been added for all of systemd's PolicyKit
2162 authenticated method calls has been added. In particular
2163 this now allows optional interactive authorization via
2164 PolicyKit for many of PID1's privileged operations such as
2165 unit file enabling and disabling.
2167 * "udevadm hwdb --update" learnt a new switch "--usr" for
2168 placing the rebuilt hardware database in /usr instead of
2169 /etc. When used only hardware database entries stored in
2170 /usr will be used, and any user database entries in /etc are
2171 ignored. This functionality is useful for vendors to ship a
2172 pre-built database on systems where local configuration is
2173 unnecessary or unlikely.
2175 * Calendar time specifications in .timer units now also
2176 understand the strings "semi-annually", "quarterly" and
2177 "minutely" as shortcuts (in addition to the preexisting
2178 "anually", "hourly", ...).
2180 * systemd-tmpfiles will now correctly create files in /dev
2181 at boot which are marked for creation only at boot. It is
2182 recommended to always create static device nodes with 'c!'
2183 and 'b!', so that they are created only at boot and not
2184 overwritten at runtime.
2186 * When the watchdog logic is used for a service (WatchdogSec=)
2187 and the watchdog timeout is hit the service will now be
2188 terminated with SIGABRT (instead of just SIGTERM), in order
2189 to make sure a proper coredump and backtrace is
2190 generated. This ensures that hanging services will result in
2191 similar coredump/backtrace behaviour as services that hit a
2194 Contributions from: Andreas Henriksson, Andrei Borzenkov,
2195 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
2196 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
2197 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
2198 Herrmann, David Sommerseth, David Strauss, Emil Renner
2199 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
2200 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
2201 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
2202 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
2203 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
2204 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
2205 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
2206 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
2207 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
2208 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
2209 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
2210 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
2211 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
2212 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
2213 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
2214 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
2217 — Berlin, 2014-10-28
2221 * timedated no longer reads NTP implementation unit names from
2222 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
2223 implementations should add a
2225 Conflicts=systemd-timesyncd.service
2227 to their unit files to take over and replace systemd's NTP
2228 default functionality.
2230 * systemd-sysusers gained a new line type "r" for configuring
2231 which UID/GID ranges to allocate system users/groups
2232 from. Lines of type "u" may now add an additional column
2233 that specifies the home directory for the system user to be
2234 created. Also, systemd-sysusers may now optionally read user
2235 information from STDIN instead of a file. This is useful for
2236 invoking it from RPM preinst scriptlets that need to create
2237 users before the first RPM file is installed since these
2238 files might need to be owned by them. A new
2239 %sysusers_create_inline RPM macro has been introduced to do
2240 just that. systemd-sysusers now updates the shadow files as
2241 well as the user/group databases, which should enhance
2242 compatibility with certain tools like grpck.
2244 * A number of bus APIs of PID 1 now optionally consult
2245 PolicyKit to permit access for otherwise unprivileged
2246 clients under certain conditions. Note that this currently
2247 doesn't support interactive authentication yet, but this is
2248 expected to be added eventually, too.
2250 * /etc/machine-info now has new fields for configuring the
2251 deployment environment of the machine, as well as the
2252 location of the machine. hostnamectl has been updated with
2253 new command to update these fields.
2255 * systemd-timesyncd has been updated to automatically acquire
2256 NTP server information from systemd-networkd, which might
2257 have been discovered via DHCP.
2259 * systemd-resolved now includes a caching DNS stub resolver
2260 and a complete LLMNR name resolution implementation. A new
2261 NSS module "nss-resolve" has been added which can be used
2262 instead of glibc's own "nss-dns" to resolve hostnames via
2263 systemd-resolved. Hostnames, addresses and arbitrary RRs may
2264 be resolved via systemd-resolved D-Bus APIs. In contrast to
2265 the glibc internal resolver systemd-resolved is aware of
2266 multi-homed system, and keeps DNS server and caches separate
2267 and per-interface. Queries are sent simultaneously on all
2268 interfaces that have DNS servers configured, in order to
2269 properly handle VPNs and local LANs which might resolve
2270 separate sets of domain names. systemd-resolved may acquire
2271 DNS server information from systemd-networkd automatically,
2272 which in turn might have discovered them via DHCP. A tool
2273 "systemd-resolve-host" has been added that may be used to
2274 query the DNS logic in resolved. systemd-resolved implements
2275 IDNA and automatically uses IDNA or UTF-8 encoding depending
2276 on whether classic DNS or LLMNR is used as transport. In the
2277 next releases we intend to add a DNSSEC and mDNS/DNS-SD
2278 implementation to systemd-resolved.
2280 * A new NSS module nss-mymachines has been added, that
2281 automatically resolves the names of all local registered
2282 containers to their respective IP addresses.
2284 * A new client tool "networkctl" for systemd-networkd has been
2285 added. It currently is entirely passive and will query
2286 networking configuration from udev, rtnetlink and networkd,
2287 and present it to the user in a very friendly
2288 way. Eventually, we hope to extend it to become a full
2289 control utility for networkd.
2291 * .socket units gained a new DeferAcceptSec= setting that
2292 controls the kernels' TCP_DEFER_ACCEPT sockopt for
2293 TCP. Similarly, support for controlling TCP keep-alive
2294 settings has been added (KeepAliveTimeSec=,
2295 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
2296 turning off Nagle's algorithm on TCP has been added
2299 * logind learned a new session type "web", for use in projects
2300 like Cockpit which register web clients as PAM sessions.
2302 * timer units with at least one OnCalendar= setting will now
2303 be started only after timer-sync.target has been
2304 reached. This way they will not elapse before the system
2305 clock has been corrected by a local NTP client or
2306 similar. This is particular useful on RTC-less embedded
2307 machines, that come up with an invalid system clock.
2309 * systemd-nspawn's --network-veth= switch should now result in
2310 stable MAC addresses for both the outer and the inner side
2313 * systemd-nspawn gained a new --volatile= switch for running
2314 container instances with /etc or /var unpopulated.
2316 * The kdbus client code has been updated to use the new Linux
2317 3.17 memfd subsystem instead of the old kdbus-specific one.
2319 * systemd-networkd's DHCP client and server now support
2320 FORCERENEW. There are also new configuration options to
2321 configure the vendor client identifier and broadcast mode
2324 * systemd will no longer inform the kernel about the current
2325 timezone, as this is necessarily incorrect and racy as the
2326 kernel has no understanding of DST and similar
2327 concepts. This hence means FAT timestamps will be always
2328 considered UTC, similar to what Android is already
2329 doing. Also, when the RTC is configured to the local time
2330 (rather than UTC) systemd will never synchronize back to it,
2331 as this might confuse Windows at a later boot.
2333 * systemd-analyze gained a new command "verify" for offline
2334 validation of unit files.
2336 * systemd-networkd gained support for a couple of additional
2337 settings for bonding networking setups. Also, the metric for
2338 statically configured routes may now be configured. For
2339 network interfaces where this is appropriate the peer IP
2340 address may now be configured.
2342 * systemd-networkd's DHCP client will no longer request
2343 broadcasting by default, as this tripped up some networks.
2344 For hardware where broadcast is required the feature should
2345 be switched back on using RequestBroadcast=yes.
2347 * systemd-networkd will now set up IPv4LL addresses (when
2348 enabled) even if DHCP is configured successfully.
2350 * udev will now default to respect network device names given
2351 by the kernel when the kernel indicates that these are
2352 predictable. This behavior can be tweaked by changing
2353 NamePolicy= in the relevant .link file.
2355 * A new library systemd-terminal has been added that
2356 implements full TTY stream parsing and rendering. This
2357 library is supposed to be used later on for implementing a
2358 full userspace VT subsystem, replacing the current kernel
2361 * A new tool systemd-journal-upload has been added to push
2362 journal data to a remote system running
2363 systemd-journal-remote.
2365 * journald will no longer forward all local data to another
2366 running syslog daemon. This change has been made because
2367 rsyslog (which appears to be the most commonly used syslog
2368 implementation these days) no longer makes use of this, and
2369 instead pulls the data out of the journal on its own. Since
2370 forwarding the messages to a non-existent syslog server is
2371 more expensive than we assumed we have now turned this
2372 off. If you run a syslog server that is not a recent rsyslog
2373 version, you have to turn this option on again
2374 (ForwardToSyslog= in journald.conf).
2376 * journald now optionally supports the LZ4 compressor for
2377 larger journal fields. This compressor should perform much
2378 better than XZ which was the previous default.
2380 * machinectl now shows the IP addresses of local containers,
2381 if it knows them, plus the interface name of the container.
2383 * A new tool "systemd-escape" has been added that makes it
2384 easy to escape strings to build unit names and similar.
2386 * sd_notify() messages may now include a new ERRNO= field
2387 which is parsed and collected by systemd and shown among the
2388 "systemctl status" output for a service.
2390 * A new component "systemd-firstboot" has been added that
2391 queries the most basic systemd information (timezone,
2392 hostname, root password) interactively on first
2393 boot. Alternatively it may also be used to provision these
2394 things offline on OS images installed into directories.
2396 * The default sysctl.d/ snippets will now set
2398 net.ipv4.conf.default.promote_secondaries=1
2400 This has the benefit of no flushing secondary IP addresses
2401 when primary addresses are removed.
2403 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
2404 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
2405 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
2406 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
2407 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
2408 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
2409 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
2410 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
2411 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
2412 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
2413 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
2414 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
2415 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
2416 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
2417 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
2419 — Berlin, 2014-08-19
2423 * A new tool systemd-sysusers has been added. This tool
2424 creates system users and groups in /etc/passwd and
2425 /etc/group, based on static declarative system user/group
2426 definitions in /usr/lib/sysusers.d/. This is useful to
2427 enable factory resets and volatile systems that boot up with
2428 an empty /etc directory, and thus need system users and
2429 groups created during early boot. systemd now also ships
2430 with two default sysusers.d/ files for the most basic
2431 users and groups systemd and the core operating system
2434 * A new tmpfiles snippet has been added that rebuilds the
2435 essential files in /etc on boot, should they be missing.
2437 * A directive for ensuring automatic clean-up of
2438 /var/cache/man/ has been removed from the default
2439 configuration. This line should now be shipped by the man
2440 implementation. The necessary change has been made to the
2441 man-db implementation. Note that you need to update your man
2442 implementation to one that ships this line, otherwise no
2443 automatic clean-up of /var/cache/man will take place.
2445 * A new condition ConditionNeedsUpdate= has been added that
2446 may conditionalize services to only run when /etc or /var
2447 are "older" than the vendor operating system resources in
2448 /usr. This is useful for reconstructing or updating /etc
2449 after an offline update of /usr or a factory reset, on the
2450 next reboot. Services that want to run once after such an
2451 update or reset should use this condition and order
2452 themselves before the new systemd-update-done.service, which
2453 will mark the two directories as fully updated. A number of
2454 service files have been added making use of this, to rebuild
2455 the udev hardware database, the journald message catalog and
2456 dynamic loader cache (ldconfig). The systemd-sysusers tool
2457 described above also makes use of this now. With this in
2458 place it is now possible to start up a minimal operating
2459 system with /etc empty cleanly. For more information on the
2460 concepts involved see this recent blog story:
2462 http://0pointer.de/blog/projects/stateless.html
2464 * A new system group "input" has been introduced, and all
2465 input device nodes get this group assigned. This is useful
2466 for system-level software to get access to input devices. It
2467 complements what is already done for "audio" and "video".
2469 * systemd-networkd learnt minimal DHCPv4 server support in
2470 addition to the existing DHCPv4 client support. It also
2471 learnt DHCPv6 client and IPv6 Router Solicitation client
2472 support. The DHCPv4 client gained support for static routes
2473 passed in from the server. Note that the [DHCPv4] section
2474 known in older systemd-networkd versions has been renamed to
2475 [DHCP] and is now also used by the DHCPv6 client. Existing
2476 .network files using settings of this section should be
2477 updated, though compatibility is maintained. Optionally, the
2478 client hostname may now be sent to the DHCP server.
2480 * networkd gained support for vxlan virtual networks as well
2481 as tun/tap and dummy devices.
2483 * networkd gained support for automatic allocation of address
2484 ranges for interfaces from a system-wide pool of
2485 addresses. This is useful for dynamically managing a large
2486 number of interfaces with a single network configuration
2487 file. In particular this is useful to easily assign
2488 appropriate IP addresses to the veth links of a large number
2489 of nspawn instances.
2491 * RPM macros for processing sysusers, sysctl and binfmt
2492 drop-in snippets at package installation time have been
2495 * The /etc/os-release file should now be placed in
2496 /usr/lib/os-release. The old location is automatically
2497 created as symlink. /usr/lib is the more appropriate
2498 location of this file, since it shall actually describe the
2499 vendor operating system shipped in /usr, and not the
2500 configuration stored in /etc.
2502 * .mount units gained a new boolean SloppyOptions= setting
2503 that maps to mount(8)'s -s option which enables permissive
2504 parsing of unknown mount options.
2506 * tmpfiles learnt a new "L+" directive which creates a symlink
2507 but (unlike "L") deletes a pre-existing file first, should
2508 it already exist and not already be the correct
2509 symlink. Similarly, "b+", "c+" and "p+" directives have been
2510 added as well, which create block and character devices, as
2511 well as fifos in the filesystem, possibly removing any
2512 pre-existing files of different types.
2514 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
2515 'argument' field (which so far specified the source to
2516 symlink/copy the files from) is now optional. If omitted the
2517 same file os copied from /usr/share/factory/ suffixed by the
2518 full destination path. This is useful for populating /etc
2519 with essential files, by copying them from vendor defaults
2520 shipped in /usr/share/factory/etc.
2522 * A new command "systemctl preset-all" has been added that
2523 applies the service preset settings to all installed unit
2524 files. A new switch --preset-mode= has been added that
2525 controls whether only enable or only disable operations
2528 * A new command "systemctl is-system-running" has been added
2529 that allows checking the overall state of the system, for
2530 example whether it is fully up and running.
2532 * When the system boots up with an empty /etc, the equivalent
2533 to "systemctl preset-all" is executed during early boot, to
2534 make sure all default services are enabled after a factory
2537 * systemd now contains a minimal preset file that enables the
2538 most basic services systemd ships by default.
2540 * Unit files' [Install] section gained a new DefaultInstance=
2541 field for defining the default instance to create if a
2542 template unit is enabled with no instance specified.
2544 * A new passive target cryptsetup-pre.target has been added
2545 that may be used by services that need to make they run and
2546 finish before the first LUKS cryptographic device is set up.
2548 * The /dev/loop-control and /dev/btrfs-control device nodes
2549 are now owned by the "disk" group by default, opening up
2550 access to this group.
2552 * systemd-coredump will now automatically generate a
2553 stack trace of all core dumps taking place on the system,
2554 based on elfutils' libdw library. This stack trace is logged
2557 * systemd-coredump may now optionally store coredumps directly
2558 on disk (in /var/lib/systemd/coredump, possibly compressed),
2559 instead of storing them unconditionally in the journal. This
2560 mode is the new default. A new configuration file
2561 /etc/systemd/coredump.conf has been added to configure this
2562 and other parameters of systemd-coredump.
2564 * coredumpctl gained a new "info" verb to show details about a
2565 specific coredump. A new switch "-1" has also been added
2566 that makes sure to only show information about the most
2567 recent entry instead of all entries. Also, as the tool is
2568 generally useful now the "systemd-" prefix of the binary
2569 name has been removed. Distributions that want to maintain
2570 compatibility with the old name should add a symlink from
2571 the old name to the new name.
2573 * journald's SplitMode= now defaults to "uid". This makes sure
2574 that unprivileged users can access their own coredumps with
2575 coredumpctl without restrictions.
2577 * New kernel command line options "systemd.wants=" (for
2578 pulling an additional unit during boot), "systemd.mask="
2579 (for masking a specific unit for the boot), and
2580 "systemd.debug-shell" (for enabling the debug shell on tty9)
2581 have been added. This is implemented in the new generator
2582 "systemd-debug-generator".
2584 * systemd-nspawn will now by default filter a couple of
2585 syscalls for containers, among them those required for
2586 kernel module loading, direct x86 IO port access, swap
2587 management, and kexec. Most importantly though
2588 open_by_handle_at() is now prohibited for containers,
2589 closing a hole similar to a recently discussed vulnerability
2590 in docker regarding access to files on file hierarchies the
2591 container should normally not have access to. Note that, for
2592 nspawn, we generally make no security claims anyway (and
2593 this is explicitly documented in the man page), so this is
2594 just a fix for one of the most obvious problems.
2596 * A new man page file-hierarchy(7) has been added that
2597 contains a minimized, modernized version of the file system
2598 layout systemd expects, similar in style to the FHS
2599 specification or hier(5). A new tool systemd-path(1) has
2600 been added to query many of these paths for the local
2603 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
2604 longer done. Since the directory now has a per-user size
2605 limit, and is cleaned on logout this appears unnecessary,
2606 in particular since this now brings the lifecycle of this
2607 directory closer in line with how IPC objects are handled.
2609 * systemd.pc now exports a number of additional directories,
2610 including $libdir (which is useful to identify the library
2611 path for the primary architecture of the system), and a
2612 couple of drop-in directories.
2614 * udev's predictable network interface names now use the dev_port
2615 sysfs attribute, introduced in linux 3.15 instead of dev_id to
2616 distinguish between ports of the same PCI function. dev_id should
2617 only be used for ports using the same HW address, hence the need
2620 * machined has been updated to export the OS version of a
2621 container (read from /etc/os-release and
2622 /usr/lib/os-release) on the bus. This is now shown in
2623 "machinectl status" for a machine.
2625 * A new service setting RestartForceExitStatus= has been
2626 added. If configured to a set of exit signals or process
2627 return values, the service will be restarted when the main
2628 daemon process exits with any of them, regardless of the
2631 * systemctl's -H switch for connecting to remote systemd
2632 machines has been extended so that it may be used to
2633 directly connect to a specific container on the
2634 host. "systemctl -H root@foobar:waldi" will now connect as
2635 user "root" to host "foobar", and then proceed directly to
2636 the container named "waldi". Note that currently you have to
2637 authenticate as user "root" for this to work, as entering
2638 containers is a privileged operation.
2640 Contributions from: Andreas Henriksson, Benjamin Steinwender,
2641 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
2642 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
2643 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
2644 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
2645 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
2646 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
2647 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
2648 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
2649 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
2650 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
2651 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
2653 — Berlin, 2014-07-03
2657 * As an experimental feature, udev now tries to lock the
2658 disk device node (flock(LOCK_SH|LOCK_NB)) while it
2659 executes events for the disk or any of its partitions.
2660 Applications like partitioning programs can lock the
2661 disk device node (flock(LOCK_EX)) and claim temporary
2662 device ownership that way; udev will entirely skip all event
2663 handling for this disk and its partitions. If the disk
2664 was opened for writing, the close will trigger a partition
2665 table rescan in udev's "watch" facility, and if needed
2666 synthesize "change" events for the disk and all its partitions.
2667 This is now unconditionally enabled, and if it turns out to
2668 cause major problems, we might turn it on only for specific
2669 devices, or might need to disable it entirely. Device Mapper
2670 devices are excluded from this logic.
2672 * We temporarily dropped the "-l" switch for fsck invocations,
2673 since they collide with the flock() logic above. util-linux
2674 upstream has been changed already to avoid this conflict,
2675 and we will readd "-l" as soon as util-linux with this
2676 change has been released.
2678 * The dependency on libattr has been removed. Since a long
2679 time, the extended attribute calls have moved to glibc, and
2680 libattr is thus unnecessary.
2682 * Virtualization detection works without priviliges now. This
2683 means the systemd-detect-virt binary no longer requires
2684 CAP_SYS_PTRACE file capabilities, and our daemons can run
2685 with fewer privileges.
2687 * systemd-networkd now runs under its own "systemd-network"
2688 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
2689 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
2690 loses the ability to write to files owned by root this way.
2692 * Similarly, systemd-resolved now runs under its own
2693 "systemd-resolve" user with no capabilities remaining.
2695 * Similarly, systemd-bus-proxyd now runs under its own
2696 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
2698 * systemd-networkd gained support for setting up "veth"
2699 virtual Ethernet devices for container connectivity, as well
2700 as GRE and VTI tunnels.
2702 * systemd-networkd will no longer automatically attempt to
2703 manually load kernel modules necessary for certain tunnel
2704 transports. Instead, it is assumed the kernel loads them
2705 automatically when required. This only works correctly on
2706 very new kernels. On older kernels, please consider adding
2707 the kernel modules to /etc/modules-load.d/ as a work-around.
2709 * The resolv.conf file systemd-resolved generates has been
2710 moved to /run/systemd/resolve/. If you have a symlink from
2711 /etc/resolv.conf, it might be necessary to correct it.
2713 * Two new service settings, ProtectHome= and ProtectSystem=,
2714 have been added. When enabled, they will make the user data
2715 (such as /home) inaccessible or read-only and the system
2716 (such as /usr) read-only, for specific services. This allows
2717 very light-weight per-service sandboxing to avoid
2718 modifications of user data or system files from
2719 services. These two new switches have been enabled for all
2720 of systemd's long-running services, where appropriate.
2722 * Socket units gained new SocketUser= and SocketGroup=
2723 settings to set the owner user and group of AF_UNIX sockets
2724 and FIFOs in the file system.
2726 * Socket units gained a new RemoveOnStop= setting. If enabled,
2727 all FIFOS and sockets in the file system will be removed
2728 when the specific socket unit is stopped.
2730 * Socket units gained a new Symlinks= setting. It takes a list
2731 of symlinks to create to file system sockets or FIFOs
2732 created by the specific Unix sockets. This is useful to
2733 manage symlinks to socket nodes with the same life-cycle as
2736 * The /dev/log socket and /dev/initctl FIFO have been moved to
2737 /run, and have been replaced by symlinks. This allows
2738 connecting to these facilities even if PrivateDevices=yes is
2739 used for a service (which makes /dev/log itself unavailable,
2740 but /run is left). This also has the benefit of ensuring
2741 that /dev only contains device nodes, directories and
2742 symlinks, and nothing else.
2744 * sd-daemon gained two new calls sd_pid_notify() and
2745 sd_pid_notifyf(). They are similar to sd_notify() and
2746 sd_notifyf(), but allow overriding of the source PID of
2747 notification messages if permissions permit this. This is
2748 useful to send notify messages on behalf of a different
2749 process (for example, the parent process). The
2750 systemd-notify tool has been updated to make use of this
2751 when sending messages (so that notification messages now
2752 originate from the shell script invoking systemd-notify and
2753 not the systemd-notify process itself. This should minimize
2754 a race where systemd fails to associate notification
2755 messages to services when the originating process already
2758 * A new "on-abnormal" setting for Restart= has been added. If
2759 set, it will result in automatic restarts on all "abnormal"
2760 reasons for a process to exit, which includes unclean
2761 signals, core dumps, timeouts and watchdog timeouts, but
2762 does not include clean and unclean exit codes or clean
2763 signals. Restart=on-abnormal is an alternative for
2764 Restart=on-failure for services that shall be able to
2765 terminate and avoid restarts on certain errors, by
2766 indicating so with an unclean exit code. Restart=on-failure
2767 or Restart=on-abnormal is now the recommended setting for
2768 all long-running services.
2770 * If the InaccessibleDirectories= service setting points to a
2771 mount point (or if there are any submounts contained within
2772 it), it is now attempted to completely unmount it, to make
2773 the file systems truly unavailable for the respective
2776 * The ReadOnlyDirectories= service setting and
2777 systemd-nspawn's --read-only parameter are now recursively
2778 applied to all submounts, too.
2780 * Mount units may now be created transiently via the bus APIs.
2782 * The support for SysV and LSB init scripts has been removed
2783 from the systemd daemon itself. Instead, it is now
2784 implemented as a generator that creates native systemd units
2785 from these scripts when needed. This enables us to remove a
2786 substantial amount of legacy code from PID 1, following the
2787 fact that many distributions only ship a very small number
2788 of LSB/SysV init scripts nowadays.
2790 * Privileged Xen (dom0) domains are not considered
2791 virtualization anymore by the virtualization detection
2792 logic. After all, they generally have unrestricted access to
2793 the hardware and usually are used to manage the unprivileged
2796 * systemd-tmpfiles gained a new "C" line type, for copying
2797 files or entire directories.
2799 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
2800 lines. So far, they have been non-globbing versions of the
2801 latter, and have thus been redundant. In future, it is
2802 recommended to only use "z". "m" has hence been removed
2803 from the documentation, even though it stays supported.
2805 * A tmpfiles snippet to recreate the most basic structure in
2806 /var has been added. This is enough to create the /var/run →
2807 /run symlink and create a couple of structural
2808 directories. This allows systems to boot up with an empty or
2809 volatile /var. Of course, while with this change, the core OS
2810 now is capable with dealing with a volatile /var, not all
2811 user services are ready for it. However, we hope that sooner
2812 or later, many service daemons will be changed upstream so
2813 that they are able to automatically create their necessary
2814 directories in /var at boot, should they be missing. This is
2815 the first step to allow state-less systems that only require
2816 the vendor image for /usr to boot.
2818 * systemd-nspawn has gained a new --tmpfs= switch to mount an
2819 empty tmpfs instance to a specific directory. This is
2820 particularly useful for making use of the automatic
2821 reconstruction of /var (see above), by passing --tmpfs=/var.
2823 * Access modes specified in tmpfiles snippets may now be
2824 prefixed with "~", which indicates that they shall be masked
2825 by whether the existing file or directory is currently
2826 writable, readable or executable at all. Also, if specified,
2827 the sgid/suid/sticky bits will be masked for all
2830 * A new passive target unit "network-pre.target" has been
2831 added which is useful for services that shall run before any
2832 network is configured, for example firewall scripts.
2834 * The "floppy" group that previously owned the /dev/fd*
2835 devices is no longer used. The "disk" group is now used
2836 instead. Distributions should probably deprecate usage of
2839 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
2840 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
2841 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
2842 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
2843 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
2844 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
2847 — Berlin, 2014-06-11
2851 * A new "systemd-timesyncd" daemon has been added for
2852 synchronizing the system clock across the network. It
2853 implements an SNTP client. In contrast to NTP
2854 implementations such as chrony or the NTP reference server,
2855 this only implements a client side, and does not bother with
2856 the full NTP complexity, focusing only on querying time from
2857 one remote server and synchronizing the local clock to
2858 it. Unless you intend to serve NTP to networked clients or
2859 want to connect to local hardware clocks, this simple NTP
2860 client should be more than appropriate for most
2861 installations. The daemon runs with minimal privileges, and
2862 has been hooked up with networkd to only operate when
2863 network connectivity is available. The daemon saves the
2864 current clock to disk every time a new NTP sync has been
2865 acquired, and uses this to possibly correct the system clock
2866 early at bootup, in order to accommodate for systems that
2867 lack an RTC such as the Raspberry Pi and embedded devices,
2868 and to make sure that time monotonically progresses on these
2869 systems, even if it is not always correct. To make use of
2870 this daemon, a new system user and group "systemd-timesync"
2871 needs to be created on installation of systemd.
2873 * The queue "seqnum" interface of libudev has been disabled, as
2874 it was generally incompatible with device namespacing as
2875 sequence numbers of devices go "missing" if the devices are
2876 part of a different namespace.
2878 * "systemctl list-timers" and "systemctl list-sockets" gained
2879 a --recursive switch for showing units of these types also
2880 for all local containers, similar in style to the already
2881 supported --recursive switch for "systemctl list-units".
2883 * A new RebootArgument= setting has been added for service
2884 units, which may be used to specify a kernel reboot argument
2885 to use when triggering reboots with StartLimitAction=.
2887 * A new FailureAction= setting has been added for service
2888 units which may be used to specify an operation to trigger
2889 when a service fails. This works similarly to
2890 StartLimitAction=, but unlike it, controls what is done
2891 immediately rather than only after several attempts to
2892 restart the service in question.
2894 * hostnamed got updated to also expose the kernel name,
2895 release, and version on the bus. This is useful for
2896 executing commands like hostnamectl with the -H switch.
2897 systemd-analyze makes use of this to properly display
2898 details when running non-locally.
2900 * The bootchart tool can now show cgroup information in the
2901 graphs it generates.
2903 * The CFS CPU quota cgroup attribute is now exposed for
2904 services. The new CPUQuota= switch has been added for this
2905 which takes a percentage value. Setting this will have the
2906 result that a service may never get more CPU time than the
2907 specified percentage, even if the machine is otherwise idle.
2909 * systemd-networkd learned IPIP and SIT tunnel support.
2911 * LSB init scripts exposing a dependency on $network will now
2912 get a dependency on network-online.target rather than simply
2913 network.target. This should bring LSB handling closer to
2914 what it was on SysV systems.
2916 * A new fsck.repair= kernel option has been added to control
2917 how fsck shall deal with unclean file systems at boot.
2919 * The (.ini) configuration file parser will now silently
2920 ignore sections whose name begins with "X-". This may be
2921 used to maintain application-specific extension sections in unit
2924 * machined gained a new API to query the IP addresses of
2925 registered containers. "machinectl status" has been updated
2926 to show these addresses in its output.
2928 * A new call sd_uid_get_display() has been added to the
2929 sd-login APIs for querying the "primary" session of a
2930 user. The "primary" session of the user is elected from the
2931 user's sessions and generally a graphical session is
2932 preferred over a text one.
2934 * A minimal systemd-resolved daemon has been added. It
2935 currently simply acts as a companion to systemd-networkd and
2936 manages resolv.conf based on per-interface DNS
2937 configuration, possibly supplied via DHCP. In the long run
2938 we hope to extend this into a local DNSSEC enabled DNS and
2941 * The systemd-networkd-wait-online tool is now enabled by
2942 default. It will delay network-online.target until a network
2943 connection has been configured. The tool primarily integrates
2944 with networkd, but will also make a best effort to make sense
2945 of network configuration performed in some other way.
2947 * Two new service options StartupCPUShares= and
2948 StartupBlockIOWeight= have been added that work similarly to
2949 CPUShares= and BlockIOWeight= however only apply during
2950 system startup. This is useful to prioritize certain services
2951 differently during bootup than during normal runtime.
2953 * hostnamed has been changed to prefer the statically
2954 configured hostname in /etc/hostname (unless set to
2955 'localhost' or empty) over any dynamic one supplied by
2956 dhcp. With this change, the rules for picking the hostname
2957 match more closely the rules of other configuration settings
2958 where the local administrator's configuration in /etc always
2959 overrides any other settings.
2961 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
2962 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
2963 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
2964 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
2965 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
2966 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
2967 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
2968 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
2969 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
2970 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
2971 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
2972 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
2973 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
2974 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
2975 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
2976 Lindskog, WaLyong Cho, Will Woods, Zbigniew
2979 — Beijing, 2014-05-28
2983 * When restoring the screen brightness at boot, stay away from
2984 the darkest setting or from the lowest 5% of the available
2985 range, depending on which is the larger value of both. This
2986 should effectively protect the user from rebooting into a
2987 black screen, should the brightness have been set to minimum
2990 * sd-login gained a new sd_machine_get_class() call to
2991 determine the class ("vm" or "container") of a machine
2992 registered with machined.
2994 * sd-login gained new calls
2995 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
2996 to query the identity of the peer of a local AF_UNIX
2997 connection. They operate similarly to their sd_pid_get_xyz()
3000 * PID 1 will now maintain a system-wide system state engine
3001 with the states "starting", "running", "degraded",
3002 "maintenance", "stopping". These states are bound to system
3003 startup, normal runtime, runtime with at least one failed
3004 service, rescue/emergency mode and system shutdown. This
3005 state is shown in the "systemctl status" output when no unit
3006 name is passed. It is useful to determine system state, in
3007 particularly when doing so for many systems or containers at
3010 * A new command "list-machines" has been added to "systemctl"
3011 that lists all local OS containers and shows their system
3012 state (see above), if systemd runs inside of them.
3014 * systemctl gained a new "-r" switch to recursively enumerate
3015 units on all local containers, when used with the
3016 "list-unit" command (which is the default one that is
3017 executed when no parameters are specified).
3019 * The GPT automatic partition discovery logic will now honour
3020 two GPT partition flags: one may be set on a partition to
3021 cause it to be mounted read-only, and the other may be set
3022 on a partition to ignore it during automatic discovery.
3024 * Two new GPT type UUIDs have been added for automatic root
3025 partition discovery, for 32-bit and 64-bit ARM. This is not
3026 particularly useful for discovering the root directory on
3027 these architectures during bare-metal boots (since UEFI is
3028 not common there), but still very useful to allow booting of
3029 ARM disk images in nspawn with the -i option.
3031 * MAC addresses of interfaces created with nspawn's
3032 --network-interface= switch will now be generated from the
3033 machine name, and thus be stable between multiple invocations
3036 * logind will now automatically remove all IPC objects owned
3037 by a user if she or he fully logs out. This makes sure that
3038 users who are logged out cannot continue to consume IPC
3039 resources. This covers SysV memory, semaphores and message
3040 queues as well as POSIX shared memory and message
3041 queues. Traditionally, SysV and POSIX IPC had no life-cycle
3042 limits. With this functionality, that is corrected. This may
3043 be turned off by using the RemoveIPC= switch of logind.conf.
3045 * The systemd-machine-id-setup and tmpfiles tools gained a
3046 --root= switch to operate on a specific root directory,
3049 * journald can now forward logged messages to the TTYs of all
3050 logged in users ("wall"). This is the default for all
3051 emergency messages now.
3053 * A new tool systemd-journal-remote has been added to stream
3054 journal log messages across the network.
3056 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
3057 controller trees are mounted into it. Note that the
3058 directories mounted beneath it are not read-only. This is a
3059 security measure and is particularly useful because glibc
3060 actually includes a search logic to pick any tmpfs it can
3061 find to implement shm_open() if /dev/shm is not available
3062 (which it might very well be in namespaced setups).
3064 * machinectl gained a new "poweroff" command to cleanly power
3065 down a local OS container.
3067 * The PrivateDevices= unit file setting will now also drop the
3068 CAP_MKNOD capability from the capability bound set, and
3069 imply DevicePolicy=closed.
3071 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
3072 comprehensively on all long-running systemd services where
3073 this is appropriate.
3075 * systemd-udevd will now run in a disassociated mount
3076 namespace. To mount directories from udev rules, make sure to
3077 pull in mount units via SYSTEMD_WANTS properties.
3079 * The kdbus support gained support for uploading policy into
3080 the kernel. sd-bus gained support for creating "monitoring"
3081 connections that can eavesdrop into all bus communication
3082 for debugging purposes.
3084 * Timestamps may now be specified in seconds since the UNIX
3085 epoch Jan 1st, 1970 by specifying "@" followed by the value
3088 * Native tcpwrap support in systemd has been removed. tcpwrap
3089 is old code, not really maintained anymore and has serious
3090 shortcomings, and better options such as firewalls
3091 exist. For setups that require tcpwrap usage, please
3092 consider invoking your socket-activated service via tcpd,
3093 like on traditional inetd.
3095 * A new system.conf configuration option
3096 DefaultTimerAccuracySec= has been added that controls the
3097 default AccuracySec= setting of .timer units.
3099 * Timer units gained a new WakeSystem= switch. If enabled,
3100 timers configured this way will cause the system to resume
3101 from system suspend (if the system supports that, which most
3104 * Timer units gained a new Persistent= switch. If enabled,
3105 timers configured this way will save to disk when they have
3106 been last triggered. This information is then used on next
3107 reboot to possible execute overdue timer events, that
3108 could not take place because the system was powered off.
3109 This enables simple anacron-like behaviour for timer units.
3111 * systemctl's "list-timers" will now also list the time a
3112 timer unit was last triggered in addition to the next time
3113 it will be triggered.
3115 * systemd-networkd will now assign predictable IPv4LL
3116 addresses to its local interfaces.
3118 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
3119 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
3120 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
3121 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
3122 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
3123 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
3124 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
3125 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
3128 — Berlin, 2014-03-25
3132 * A new unit file setting RestrictAddressFamilies= has been
3133 added to restrict which socket address families unit
3134 processes gain access to. This takes address family names
3135 like "AF_INET" or "AF_UNIX", and is useful to minimize the
3136 attack surface of services via exotic protocol stacks. This
3137 is built on seccomp system call filters.
3139 * Two new unit file settings RuntimeDirectory= and
3140 RuntimeDirectoryMode= have been added that may be used to
3141 manage a per-daemon runtime directories below /run. This is
3142 an alternative for setting up directory permissions with
3143 tmpfiles snippets, and has the advantage that the runtime
3144 directory's lifetime is bound to the daemon runtime and that
3145 the daemon starts up with an empty directory each time. This
3146 is particularly useful when writing services that drop
3147 privileges using the User= or Group= setting.
3149 * The DeviceAllow= unit setting now supports globbing for
3150 matching against device group names.
3152 * The systemd configuration file system.conf gained new
3153 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
3154 DefaultMemoryAccounting= to globally turn on/off accounting
3155 for specific resources (cgroups) for all units. These
3156 settings may still be overridden individually in each unit
3159 * systemd-gpt-auto-generator is now able to discover /srv and
3160 root partitions in addition to /home and swap partitions. It
3161 also supports LUKS-encrypted partitions now. With this in
3162 place, automatic discovery of partitions to mount following
3163 the Discoverable Partitions Specification
3164 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
3165 is now a lot more complete. This allows booting without
3166 /etc/fstab and without root= on the kernel command line on
3167 systems prepared appropriately.
3169 * systemd-nspawn gained a new --image= switch which allows
3170 booting up disk images and Linux installations on any block
3171 device that follow the Discoverable Partitions Specification
3172 (see above). This means that installations made with
3173 appropriately updated installers may now be started and
3174 deployed using container managers, completely
3175 unmodified. (We hope that libvirt-lxc will add support for
3176 this feature soon, too.)
3178 * systemd-nspawn gained a new --network-macvlan= setting to
3179 set up a private macvlan interface for the
3180 container. Similarly, systemd-networkd gained a new
3181 Kind=macvlan setting in .netdev files.
3183 * systemd-networkd now supports configuring local addresses
3186 * A new tool systemd-network-wait-online has been added to
3187 synchronously wait for network connectivity using
3190 * The sd-bus.h bus API gained a new sd_bus_track object for
3191 tracking the life-cycle of bus peers. Note that sd-bus.h is
3192 still not a public API though (unless you specify
3193 --enable-kdbus on the configure command line, which however
3194 voids your warranty and you get no API stability guarantee).
3196 * The $XDG_RUNTIME_DIR runtime directories for each user are
3197 now individual tmpfs instances, which has the benefit of
3198 introducing separate pools for each user, with individual
3199 size limits, and thus making sure that unprivileged clients
3200 can no longer negatively impact the system or other users by
3201 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
3202 RuntimeDirectorySize= has been introduced that allows
3203 controlling the default size limit for all users. It
3204 defaults to 10% of the available physical memory. This is no
3205 replacement for quotas on tmpfs though (which the kernel
3206 still does not support), as /dev/shm and /tmp are still
3207 shared resources used by both the system and unprivileged
3210 * logind will now automatically turn off automatic suspending
3211 on laptop lid close when more than one display is
3212 connected. This was previously expected to be implemented
3213 individually in desktop environments (such as GNOME),
3214 however has been added to logind now, in order to fix a
3215 boot-time race where a desktop environment might not have
3216 been started yet and thus not been able to take an inhibitor
3217 lock at the time where logind already suspends the system
3218 due to a closed lid.
3220 * logind will now wait at least 30s after each system
3221 suspend/resume cycle, and 3min after system boot before
3222 suspending the system due to a closed laptop lid. This
3223 should give USB docking stations and similar enough time to
3224 be probed and configured after system resume and boot in
3225 order to then act as suspend blocker.
3227 * systemd-run gained a new --property= setting which allows
3228 initialization of resource control properties (and others)
3229 for the created scope or service unit. Example: "systemd-run
3230 --property=BlockIOWeight=10 updatedb" may be used to run
3231 updatedb at a low block IO scheduling weight.
3233 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
3234 now also work in --scope mode.
3236 * When systemd is compiled with kdbus support, basic support
3237 for enforced policies is now in place. (Note that enabling
3238 kdbus still voids your warranty and no API compatibility
3241 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
3242 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
3243 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
3244 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
3245 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
3246 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
3247 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
3248 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
3249 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
3250 Zbigniew Jędrzejewski-Szmek
3252 — Berlin, 2014-03-12
3256 * systemd will now relabel /dev after loading the SMACK policy
3257 according to SMACK rules.
3259 * A new unit file option AppArmorProfile= has been added to
3260 set the AppArmor profile for the processes of a unit.
3262 * A new condition check ConditionArchitecture= has been added
3263 to conditionalize units based on the system architecture, as
3264 reported by uname()'s "machine" field.
3266 * systemd-networkd now supports matching on the system
3267 virtualization, architecture, kernel command line, host name
3270 * logind is now a lot more aggressive when suspending the
3271 machine due to a closed laptop lid. Instead of acting only
3272 on the lid close action, it will continuously watch the lid
3273 status and act on it. This is useful for laptops where the
3274 power button is on the outside of the chassis so that it can
3275 be reached without opening the lid (such as the Lenovo
3276 Yoga). On those machines, logind will now immediately
3277 re-suspend the machine if the power button has been
3278 accidentally pressed while the laptop was suspended and in a
3279 backpack or similar.
3281 * logind will now watch SW_DOCK switches and inhibit reaction
3282 to the lid switch if it is pressed. This means that logind
3283 will not suspend the machine anymore if the lid is closed
3284 and the system is docked, if the laptop supports SW_DOCK
3285 notifications via the input layer. Note that ACPI docking
3286 stations do not generate this currently. Also note that this
3287 logic is usually not fully sufficient and Desktop
3288 Environments should take a lid switch inhibitor lock when an
3289 external display is connected, as systemd will not watch
3292 * nspawn will now make use of the devices cgroup controller by
3293 default, and only permit creation of and access to the usual
3294 API device nodes like /dev/null or /dev/random, as well as
3295 access to (but not creation of) the pty devices.
3297 * We will now ship a default .network file for
3298 systemd-networkd that automatically configures DHCP for
3299 network interfaces created by nspawn's --network-veth or
3300 --network-bridge= switches.
3302 * systemd will now understand the usual M, K, G, T suffixes
3303 according to SI conventions (i.e. to the base 1000) when
3304 referring to throughput and hardware metrics. It will stay
3305 with IEC conventions (i.e. to the base 1024) for software
3306 metrics, according to what is customary according to
3307 Wikipedia. We explicitly document which base applies for
3308 each configuration option.
3310 * The DeviceAllow= setting in unit files now supports a syntax
3311 to whitelist an entire group of devices node majors at once,
3312 based on the /proc/devices listing. For example, with the
3313 string "char-pts", it is now possible to whitelist all
3314 current and future pseudo-TTYs at once.
3316 * sd-event learned a new "post" event source. Event sources of
3317 this type are triggered by the dispatching of any event
3318 source of a type that is not "post". This is useful for
3319 implementing clean-up and check event sources that are
3320 triggered by other work being done in the program.
3322 * systemd-networkd is no longer statically enabled, but uses
3323 the usual [Install] sections so that it can be
3324 enabled/disabled using systemctl. It still is enabled by
3327 * When creating a veth interface pair with systemd-nspawn, the
3328 host side will now be prefixed with "vb-" if
3329 --network-bridge= is used, and with "ve-" if --network-veth
3330 is used. This way, it is easy to distinguish these cases on
3331 the host, for example to apply different configuration to
3332 them with systemd-networkd.
3334 * The compatibility libraries for libsystemd-journal.so,
3335 libsystem-id128.so, libsystemd-login.so and
3336 libsystemd-daemon.so do not make use of IFUNC
3337 anymore. Instead, we now build libsystemd.so multiple times
3338 under these alternative names. This means that the footprint
3339 is drastically increased, but given that these are
3340 transitional compatibility libraries, this should not matter
3341 much. This change has been made necessary to support the ARM
3342 platform for these compatibility libraries, as the ARM
3343 toolchain is not really at the same level as the toolchain
3344 for other architectures like x86 and does not support
3345 IFUNC. Please make sure to use --enable-compat-libs only
3346 during a transitional period!
3348 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
3349 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
3350 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
3351 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
3352 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
3353 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
3354 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
3355 Zbigniew Jędrzejewski-Szmek
3357 — Berlin, 2014-02-24
3361 * A new component "systemd-networkd" has been added that can
3362 be used to configure local network interfaces statically or
3363 via DHCP. It is capable of bringing up bridges, VLANs, and
3364 bonding. Currently, no hook-ups for interactive network
3365 configuration are provided. Use this for your initrd,
3366 container, embedded, or server setup if you need a simple,
3367 yet powerful, network configuration solution. This
3368 configuration subsystem is quite nifty, as it allows wildcard
3369 hotplug matching in interfaces. For example, with a single
3370 configuration snippet, you can configure that all Ethernet
3371 interfaces showing up are automatically added to a bridge,
3372 or similar. It supports link-sensing and more.
3374 * A new tool "systemd-socket-proxyd" has been added which can
3375 act as a bidirectional proxy for TCP sockets. This is
3376 useful for adding socket activation support to services that
3377 do not actually support socket activation, including virtual
3378 machines and the like.
3380 * Add a new tool to save/restore rfkill state on
3383 * Save/restore state of keyboard backlights in addition to
3384 display backlights on shutdown/boot.
3386 * udev learned a new SECLABEL{} construct to label device
3387 nodes with a specific security label when they appear. For
3388 now, only SECLABEL{selinux} is supported, but the syntax is
3389 prepared for additional security frameworks.
3391 * udev gained a new scheme to configure link-level attributes
3392 from files in /etc/systemd/network/*.link. These files can
3393 match against MAC address, device path, driver name and type,
3394 and will apply attributes like the naming policy, link speed,
3395 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
3396 address assignment policy (randomized, ...).
3398 * The configuration of network interface naming rules for
3399 "permanent interface names" has changed: a new NamePolicy=
3400 setting in the [Link] section of .link files determines the
3401 priority of possible naming schemes (onboard, slot, MAC,
3402 path). The default value of this setting is determined by
3403 /usr/lib/net/links/99-default.link. Old
3404 80-net-name-slot.rules udev configuration file has been
3405 removed, so local configuration overriding this file should
3406 be adapated to override 99-default.link instead.
3408 * When the User= switch is used in a unit file, also
3409 initialize $SHELL= based on the user database entry.
3411 * systemd no longer depends on libdbus. All communication is
3412 now done with sd-bus, systemd's low-level bus library
3415 * kdbus support has been added to PID 1 itself. When kdbus is
3416 enabled, this causes PID 1 to set up the system bus and
3417 enable support for a new ".busname" unit type that
3418 encapsulates bus name activation on kdbus. It works a little
3419 bit like ".socket" units, except for bus names. A new
3420 generator has been added that converts classic dbus1 service
3421 activation files automatically into native systemd .busname
3424 * sd-bus: add a light-weight vtable implementation that allows
3425 defining objects on the bus with a simple static const
3426 vtable array of its methods, signals and properties.
3428 * systemd will not generate or install static dbus
3429 introspection data anymore to /usr/share/dbus-1/interfaces,
3430 as the precise format of these files is unclear, and
3431 nothing makes use of it.
3433 * A proxy daemon is now provided to proxy clients connecting
3434 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
3435 compatibility with classic D-Bus.
3437 * A bus driver implementation has been added that supports the
3438 classic D-Bus bus driver calls on kdbus, also for
3439 compatibility purposes.
3441 * A new API "sd-event.h" has been added that implements a
3442 minimal event loop API built around epoll. It provides a
3443 couple of features that direct epoll usage is lacking:
3444 prioritization of events, scales to large numbers of timer
3445 events, per-event timer slack (accuracy), system-wide
3446 coalescing of timer events, exit handlers, watchdog
3447 supervision support using systemd's sd_notify() API, child
3450 * A new API "sd-rntl.h" has been added that provides an API
3451 around the route netlink interface of the kernel, similar in
3452 style to "sd-bus.h".
3454 * A new API "sd-dhcp-client.h" has been added that provides a
3455 small DHCPv4 client-side implementation. This is used by
3458 * There is a new kernel command line option
3459 "systemd.restore_state=0|1". When set to "0", none of the
3460 systemd tools will restore saved runtime state to hardware
3461 devices. More specifically, the rfkill and backlight states
3464 * The FsckPassNo= compatibility option in mount/service units
3465 has been removed. The fstab generator will now add the
3466 necessary dependencies automatically, and does not require
3467 PID1's support for that anymore.
3469 * journalctl gained a new switch, --list-boots, that lists
3470 recent boots with their times and boot IDs.
3472 * The various tools like systemctl, loginctl, timedatectl,
3473 busctl, systemd-run, ... have gained a new switch "-M" to
3474 connect to a specific, local OS container (as direct
3475 connection, without requiring SSH). This works on any
3476 container that is registered with machined, such as those
3477 created by libvirt-lxc or nspawn.
3479 * systemd-run and systemd-analyze also gained support for "-H"
3480 to connect to remote hosts via SSH. This is particularly
3481 useful for systemd-run because it enables queuing of jobs
3482 onto remote systems.
3484 * machinectl gained a new command "login" to open a getty
3485 login in any local container. This works with any container
3486 that is registered with machined (such as those created by
3487 libvirt-lxc or nspawn), and which runs systemd inside.
3489 * machinectl gained a new "reboot" command that may be used to
3490 trigger a reboot on a specific container that is registered
3491 with machined. This works on any container that runs an init
3492 system of some kind.
3494 * systemctl gained a new "list-timers" command to print a nice
3495 listing of installed timer units with the times they elapse
3498 * Alternative reboot() parameters may now be specified on the
3499 "systemctl reboot" command line and are passed to the
3500 reboot() system call.
3502 * systemctl gained a new --job-mode= switch to configure the
3503 mode to queue a job with. This is a more generic version of
3504 --fail, --irreversible, and --ignore-dependencies, which are
3505 still available but not advertised anymore.
3507 * /etc/systemd/system.conf gained new settings to configure
3508 various default timeouts of units, as well as the default
3509 start limit interval and burst. These may still be overridden
3512 * PID1 will now export on the bus profile data of the security
3513 policy upload process (such as the SELinux policy upload to
3516 * journald: when forwarding logs to the console, include
3517 timestamps (following the setting in
3518 /sys/module/printk/parameters/time).
3520 * OnCalendar= in timer units now understands the special
3521 strings "yearly" and "annually". (Both are equivalent)
3523 * The accuracy of timer units is now configurable with the new
3524 AccuracySec= setting. It defaults to 1min.
3526 * A new dependency type JoinsNamespaceOf= has been added that
3527 allows running two services within the same /tmp and network
3528 namespace, if PrivateNetwork= or PrivateTmp= are used.
3530 * A new command "cat" has been added to systemctl. It outputs
3531 the original unit file of a unit, and concatenates the
3532 contents of additional "drop-in" unit file snippets, so that
3533 the full configuration is shown.
3535 * systemctl now supports globbing on the various "list-xyz"
3536 commands, like "list-units" or "list-sockets", as well as on
3537 those commands which take multiple unit names.
3539 * journalctl's --unit= switch gained support for globbing.
3541 * All systemd daemons now make use of the watchdog logic so
3542 that systemd automatically notices when they hang.
3544 * If the $container_ttys environment variable is set,
3545 getty-generator will automatically spawn a getty for each
3546 listed tty. This is useful for container managers to request
3547 login gettys to be spawned on as many ttys as needed.
3549 * %h, %s, %U specifier support is not available anymore when
3550 used in unit files for PID 1. This is because NSS calls are
3551 not safe from PID 1. They stay available for --user
3552 instances of systemd, and as special case for the root user.
3554 * loginctl gained a new "--no-legend" switch to turn off output
3557 * The "sd-login.h" API gained three new calls:
3558 sd_session_is_remote(), sd_session_get_remote_user(),
3559 sd_session_get_remote_host() to query information about
3562 * The udev hardware database now also carries vendor/product
3563 information of SDIO devices.
3565 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
3566 determine whether watchdog notifications are requested by
3569 * Socket-activated per-connection services now include a
3570 short description of the connection parameters in the
3573 * tmpfiles gained a new "--boot" option. When this is not used,
3574 only lines where the command character is not suffixed with
3575 "!" are executed. When this option is specified, those
3576 options are executed too. This partitions tmpfiles
3577 directives into those that can be safely executed at any
3578 time, and those which should be run only at boot (for
3579 example, a line that creates /run/nologin).
3581 * A new API "sd-resolve.h" has been added which provides a simple
3582 asynchronous wrapper around glibc NSS host name resolution
3583 calls, such as getaddrinfo(). In contrast to glibc's
3584 getaddrinfo_a(), it does not use signals. In contrast to most
3585 other asynchronous name resolution libraries, this one does
3586 not reimplement DNS, but reuses NSS, so that alternate
3587 host name resolution systems continue to work, such as mDNS,
3588 LDAP, etc. This API is based on libasyncns, but it has been
3589 cleaned up for inclusion in systemd.
3591 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
3592 "sd-daemon.h" are no longer found in individual libraries
3593 libsystemd-journal.so, libsystemd-login.so,
3594 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
3595 merged them into a single library, libsystemd.so, which
3596 provides all symbols. The reason for this is cyclic
3597 dependencies, as these libraries tend to use each other's
3598 symbols. So far, we have managed to workaround that by linking
3599 a copy of a good part of our code into each of these
3600 libraries again and again, which, however, makes certain
3601 things hard to do, like sharing static variables. Also, it
3602 substantially increases footprint. With this change, there
3603 is only one library for the basic APIs systemd
3604 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
3605 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
3606 library as well, however are subject to the --enable-kdbus
3607 switch (see below). Note that "sd-dhcp-client.h" is not part
3608 of this library (this is because it only consumes, never
3609 provides, services of/to other APIs). To make the transition
3610 easy from the separate libraries to the unified one, we
3611 provide the --enable-compat-libs compile-time switch which
3612 will generate stub libraries that are compatible with the
3613 old ones but redirect all calls to the new one.
3615 * All of the kdbus logic and the new APIs "sd-bus.h",
3616 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
3617 and "sd-utf8.h" are compile-time optional via the
3618 "--enable-kdbus" switch, and they are not compiled in by
3619 default. To make use of kdbus, you have to explicitly enable
3620 the switch. Note however, that neither the kernel nor the
3621 userspace API for all of this is considered stable yet. We
3622 want to maintain the freedom to still change the APIs for
3623 now. By specifying this build-time switch, you acknowledge
3624 that you are aware of the instability of the current
3627 * Also, note that while kdbus is pretty much complete,
3628 it lacks one thing: proper policy support. This means you
3629 can build a fully working system with all features; however,
3630 it will be highly insecure. Policy support will be added in
3631 one of the next releases, at the same time that we will
3632 declare the APIs stable.
3634 * When the kernel command line argument "kdbus" is specified,
3635 systemd will automatically load the kdbus.ko kernel module. At
3636 this stage of development, it is only useful for testing kdbus
3637 and should not be used in production. Note: if "--enable-kdbus"
3638 is specified, and the kdbus.ko kernel module is available, and
3639 "kdbus" is added to the kernel command line, the entire system
3640 runs with kdbus instead of dbus-daemon, with the above mentioned
3641 problem of missing the system policy enforcement. Also a future
3642 version of kdbus.ko or a newer systemd will not be compatible with
3643 each other, and will unlikely be able to boot the machine if only
3644 one of them is updated.
3646 * systemctl gained a new "import-environment" command which
3647 uploads the caller's environment (or parts thereof) into the
3648 service manager so that it is inherited by services started
3649 by the manager. This is useful to upload variables like
3650 $DISPLAY into the user service manager.
3652 * A new PrivateDevices= switch has been added to service units
3653 which allows running a service with a namespaced /dev
3654 directory that does not contain any device nodes for
3655 physical devices. More specifically, it only includes devices
3656 such as /dev/null, /dev/urandom, and /dev/zero which are API
3659 * logind has been extended to support behaviour like VT
3660 switching on seats that do not support a VT. This makes
3661 multi-session available on seats that are not the first seat
3662 (seat0), and on systems where kernel support for VTs has
3663 been disabled at compile-time.
3665 * If a process holds a delay lock for system sleep or shutdown
3666 and fails to release it in time, we will now log its
3667 identity. This makes it easier to identify processes that
3668 cause slow suspends or power-offs.
3670 * When parsing /etc/crypttab, support for a new key-slot=
3671 option as supported by Debian is added. It allows indicating
3672 which LUKS slot to use on disk, speeding up key loading.
3674 * The sd_journald_sendv() API call has been checked and
3675 officially declared to be async-signal-safe so that it may
3676 be invoked from signal handlers for logging purposes.
3678 * Boot-time status output is now enabled automatically after a
3679 short timeout if boot does not progress, in order to give
3680 the user an indication what she or he is waiting for.
3682 * The boot-time output has been improved to show how much time
3683 remains until jobs expire.
3685 * The KillMode= switch in service units gained a new possible
3686 value "mixed". If set, and the unit is shut down, then the
3687 initial SIGTERM signal is sent only to the main daemon
3688 process, while the following SIGKILL signal is sent to
3689 all remaining processes of the service.
3691 * When a scope unit is registered, a new property "Controller"
3692 may be set. If set to a valid bus name, systemd will send a
3693 RequestStop() signal to this name when it would like to shut
3694 down the scope. This may be used to hook manager logic into
3695 the shutdown logic of scope units. Also, scope units may now
3696 be put in a special "abandoned" state, in which case the
3697 manager process which created them takes no further
3698 responsibilities for it.
3700 * When reading unit files, systemd will now verify
3701 the access mode of these files, and warn about certain
3702 suspicious combinations. This has been added to make it
3703 easier to track down packaging bugs where unit files are
3704 marked executable or world-writable.
3706 * systemd-nspawn gained a new "--setenv=" switch to set
3707 container-wide environment variables. The similar option in
3708 systemd-activate was renamed from "--environment=" to
3709 "--setenv=" for consistency.
3711 * systemd-nspawn has been updated to create a new kdbus domain
3712 for each container that is invoked, thus allowing each
3713 container to have its own set of system and user buses,
3714 independent of the host.
3716 * systemd-nspawn gained a new --drop-capability= switch to run
3717 the container with less capabilities than the default. Both
3718 --drop-capability= and --capability= now take the special
3719 string "all" for dropping or keeping all capabilities.
3721 * systemd-nspawn gained new switches for executing containers
3722 with specific SELinux labels set.
3724 * systemd-nspawn gained a new --quiet switch to not generate
3725 any additional output but the container's own console
3728 * systemd-nspawn gained a new --share-system switch to run a
3729 container without PID namespacing enabled.
3731 * systemd-nspawn gained a new --register= switch to control
3732 whether the container is registered with systemd-machined or
3733 not. This is useful for containers that do not run full
3734 OS images, but only specific apps.
3736 * systemd-nspawn gained a new --keep-unit which may be used
3737 when invoked as the only program from a service unit, and
3738 results in registration of the unit service itself in
3739 systemd-machined, instead of a newly opened scope unit.
3741 * systemd-nspawn gained a new --network-interface= switch for
3742 moving arbitrary interfaces to the container. The new
3743 --network-veth switch creates a virtual Ethernet connection
3744 between host and container. The new --network-bridge=
3745 switch then allows assigning the host side of this virtual
3746 Ethernet connection to a bridge device.
3748 * systemd-nspawn gained a new --personality= switch for
3749 setting the kernel personality for the container. This is
3750 useful when running a 32-bit container on a 64-bit host. A
3751 similar option Personality= is now also available for service
3754 * logind will now also track a "Desktop" identifier for each
3755 session which encodes the desktop environment of it. This is
3756 useful for desktop environments that want to identify
3757 multiple running sessions of itself easily.
3759 * A new SELinuxContext= setting for service units has been
3760 added that allows setting a specific SELinux execution
3761 context for a service.
3763 * Most systemd client tools will now honour $SYSTEMD_LESS for
3764 settings of the "less" pager. By default, these tools will
3765 override $LESS to allow certain operations to work, such as
3766 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
3767 influence this logic.
3769 * systemd's "seccomp" hook-up has been changed to make use of
3770 the libseccomp library instead of using its own
3771 implementation. This has benefits for portability among
3774 * For usage together with SystemCallFilter=, a new
3775 SystemCallErrorNumber= setting has been introduced that
3776 allows configuration of a system error number to be returned
3777 on filtered system calls, instead of immediately killing the
3778 process. Also, SystemCallArchitectures= has been added to
3779 limit access to system calls of a particular architecture
3780 (in order to turn off support for unused secondary
3781 architectures). There is also a global
3782 SystemCallArchitectures= setting in system.conf now to turn
3783 off support for non-native system calls system-wide.
3785 * systemd requires a kernel with a working name_to_handle_at(),
3786 please see the kernel config requirements in the README file.
3788 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
3789 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
3790 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
3791 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
3792 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
3793 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
3794 Elia Pinto, Florian Weimer, George McCollister, Goffredo
3795 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
3796 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
3797 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
3798 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
3799 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
3800 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
3801 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
3802 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
3803 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
3804 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
3805 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
3806 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
3807 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
3808 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
3809 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
3810 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
3811 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
3813 — Berlin, 2014-02-20
3817 * logind has gained support for facilitating privileged input
3818 and drm device access for unprivileged clients. This work is
3819 useful to allow Wayland display servers (and similar
3820 programs, such as kmscon) to run under the user's ID and
3821 access input and drm devices which are normally
3822 protected. When this is used (and the kernel is new enough)
3823 logind will "mute" IO on the file descriptors passed to
3824 Wayland as long as it is in the background and "unmute" it
3825 if it returns into the foreground. This allows secure
3826 session switching without allowing background sessions to
3827 eavesdrop on input and display data. This also introduces
3828 session switching support if VT support is turned off in the
3829 kernel, and on seats that are not seat0.
3831 * A new kernel command line option luks.options= is understood
3832 now which allows specifying LUKS options for usage for LUKS
3833 encrypted partitions specified with luks.uuid=.
3835 * tmpfiles.d(5) snippets may now use specifier expansion in
3836 path names. More specifically %m, %b, %H, %v, are now
3837 replaced by the local machine id, boot id, hostname, and
3838 kernel version number.
3840 * A new tmpfiles.d(5) command "m" has been introduced which
3841 may be used to change the owner/group/access mode of a file
3842 or directory if it exists, but do nothing if it does not.
3844 * This release removes high-level support for the
3845 MemorySoftLimit= cgroup setting. The underlying kernel
3846 cgroup attribute memory.soft_limit= is currently badly
3847 designed and likely to be removed from the kernel API in its
3848 current form, hence we should not expose it for now.
3850 * The memory.use_hierarchy cgroup attribute is now enabled for
3851 all cgroups systemd creates in the memory cgroup
3852 hierarchy. This option is likely to be come the built-in
3853 default in the kernel anyway, and the non-hierarchical mode
3854 never made much sense in the intrinsically hierarchical
3857 * A new field _SYSTEMD_SLICE= is logged along with all journal
3858 messages containing the slice a message was generated
3859 from. This is useful to allow easy per-customer filtering of
3860 logs among other things.
3862 * systemd-journald will no longer adjust the group of journal
3863 files it creates to the "systemd-journal" group. Instead we
3864 rely on the journal directory to be owned by the
3865 "systemd-journal" group, and its setgid bit set, so that the
3866 kernel file system layer will automatically enforce that
3867 journal files inherit this group assignment. The reason for
3868 this change is that we cannot allow NSS look-ups from
3869 journald which would be necessary to resolve
3870 "systemd-journal" to a numeric GID, because this might
3871 create deadlocks if NSS involves synchronous queries to
3872 other daemons (such as nscd, or sssd) which in turn are
3873 logging clients of journald and might block on it, which
3874 would then dead lock. A tmpfiles.d(5) snippet included in
3875 systemd will make sure the setgid bit and group are
3876 properly set on the journal directory if it exists on every
3877 boot. However, we recommend adjusting it manually after
3878 upgrades too (or from RPM scriptlets), so that the change is
3879 not delayed until next reboot.
3881 * Backlight and random seed files in /var/lib/ have moved into
3882 the /var/lib/systemd/ directory, in order to centralize all
3883 systemd generated files in one directory.
3885 * Boot time performance measurements (as displayed by
3886 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
3887 performance information if that's available to determine how
3888 much time BIOS and boot loader initialization required. With
3889 a sufficiently new BIOS you hence no longer need to boot
3890 with Gummiboot to get access to such information.
3892 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
3893 Cristian Rodríguez, Dave Reisner, David Herrmann, David
3894 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
3895 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
3896 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
3897 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
3898 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
3900 — Berlin, 2013-10-02
3904 * The Restart= option for services now understands a new
3905 on-watchdog setting, which will restart the service
3906 automatically if the service stops sending out watchdog keep
3907 alive messages (as configured with WatchdogSec=).
3909 * The getty generator (which is responsible for bringing up a
3910 getty on configured serial consoles) will no longer only
3911 start a getty on the primary kernel console but on all
3912 others, too. This makes the order in which console= is
3913 specified on the kernel command line less important.
3915 * libsystemd-logind gained a new sd_session_get_vt() call to
3916 retrieve the VT number of a session.
3918 * If the option "tries=0" is set for an entry of /etc/crypttab
3919 its passphrase is queried indefinitely instead of any
3920 maximum number of tries.
3922 * If a service with a configure PID file terminates its PID
3923 file will now be removed automatically if it still exists
3924 afterwards. This should put an end to stale PID files.
3926 * systemd-run will now also take relative binary path names
3927 for execution and no longer insists on absolute paths.
3929 * InaccessibleDirectories= and ReadOnlyDirectories= now take
3930 paths that are optionally prefixed with "-" to indicate that
3931 it should not be considered a failure if they do not exist.
3933 * journalctl -o (and similar commands) now understands a new
3934 output mode "short-precise", it is similar to "short" but
3935 shows timestamps with usec accuracy.
3937 * The option "discard" (as known from Debian) is now
3938 synonymous to "allow-discards" in /etc/crypttab. In fact,
3939 "discard" is preferred now (since it is easier to remember
3942 * Some licensing clean-ups were made, so that more code is now
3943 LGPL-2.1 licensed than before.
3945 * A minimal tool to save/restore the display backlight
3946 brightness across reboots has been added. It will store the
3947 backlight setting as late as possible at shutdown, and
3948 restore it as early as possible during reboot.
3950 * A logic to automatically discover and enable home and swap
3951 partitions on GPT disks has been added. With this in place
3952 /etc/fstab becomes optional for many setups as systemd can
3953 discover certain partitions located on the root disk
3954 automatically. Home partitions are recognized under their
3955 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
3956 partitions are recognized under their GPT type ID
3957 0657fd6da4ab43c484e50933c84b4f4f.
3959 * systemd will no longer pass any environment from the kernel
3960 or initrd to system services. If you want to set an
3961 environment for all services, do so via the kernel command
3962 line systemd.setenv= assignment.
3964 * The systemd-sysctl tool no longer natively reads the file
3965 /etc/sysctl.conf. If desired, the file should be symlinked
3966 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
3967 legacy support by a symlink rather than built-in code, it
3968 also makes the otherwise hidden order of application of the
3969 different files visible. (Note that this partly reverts to a
3970 pre-198 application order of sysctl knobs!)
3972 * The "systemctl set-log-level" and "systemctl dump" commands
3973 have been moved to systemd-analyze.
3975 * systemd-run learned the new --remain-after-exit switch,
3976 which causes the scope unit not to be cleaned up
3977 automatically after the process terminated.
3979 * tmpfiles learned a new --exclude-prefix= switch to exclude
3980 certain paths from operation.
3982 * journald will now automatically flush all messages to disk
3983 as soon as a message at the log level CRIT, ALERT or EMERG
3986 Contributions from: Andrew Cook, Brandon Philips, Christian
3987 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
3988 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
3989 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
3990 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
3991 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
3992 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
3993 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
3994 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
3995 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
3996 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
3997 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
3998 William Giokas, Zbigniew Jędrzejewski-Szmek
4000 — Berlin, 2013-09-13
4004 * The documentation has been updated to cover the various new
4005 concepts introduced with 205.
4007 * Unit files now understand the new %v specifier which
4008 resolves to the kernel version string as returned by "uname
4011 * systemctl now supports filtering the unit list output by
4012 load state, active state and sub state, using the new
4015 * "systemctl status" will now show the results of the
4016 condition checks (like ConditionPathExists= and similar) of
4017 the last start attempts of the unit. They are also logged to
4020 * "journalctl -b" may now be used to look for boot output of a
4021 specific boot. Try "journalctl -b -1" for the previous boot,
4022 but the syntax is substantially more powerful.
4024 * "journalctl --show-cursor" has been added which prints the
4025 cursor string the last shown log line. This may then be used
4026 with the new "journalctl --after-cursor=" switch to continue
4027 browsing logs from that point on.
4029 * "journalctl --force" may now be used to force regeneration
4032 * Creation of "dead" device nodes has been moved from udev
4033 into kmod and tmpfiles. Previously, udev would read the kmod
4034 databases to pre-generate dead device nodes based on meta
4035 information contained in kernel modules, so that these would
4036 be auto-loaded on access rather then at boot. As this
4037 does not really have much to do with the exposing actual
4038 kernel devices to userspace this has always been slightly
4039 alien in the udev codebase. Following the new scheme kmod
4040 will now generate a runtime snippet for tmpfiles from the
4041 module meta information and it now is tmpfiles' job to the
4042 create the nodes. This also allows overriding access and
4043 other parameters for the nodes using the usual tmpfiles
4044 facilities. As side effect this allows us to remove the
4045 CAP_SYS_MKNOD capability bit from udevd entirely.
4047 * logind's device ACLs may now be applied to these "dead"
4048 devices nodes too, thus finally allowing managed access to
4049 devices such as /dev/snd/sequencer whithout loading the
4050 backing module right-away.
4052 * A new RPM macro has been added that may be used to apply
4053 tmpfiles configuration during package installation.
4055 * systemd-detect-virt and ConditionVirtualization= now can
4056 detect User-Mode-Linux machines (UML).
4058 * journald will now implicitly log the effective capabilities
4059 set of processes in the message metadata.
4061 * systemd-cryptsetup has gained support for TrueCrypt volumes.
4063 * The initrd interface has been simplified (more specifically,
4064 support for passing performance data via environment
4065 variables and fsck results via files in /run has been
4066 removed). These features were non-essential, and are
4067 nowadays available in a much nicer way by having systemd in
4068 the initrd serialize its state and have the hosts systemd
4069 deserialize it again.
4071 * The udev "keymap" data files and tools to apply keyboard
4072 specific mappings of scan to key codes, and force-release
4073 scan code lists have been entirely replaced by a udev
4074 "keyboard" builtin and a hwdb data file.
4076 * systemd will now honour the kernel's "quiet" command line
4077 argument also during late shutdown, resulting in a
4078 completely silent shutdown when used.
4080 * There's now an option to control the SO_REUSEPORT socket
4081 option in .socket units.
4083 * Instance units will now automatically get a per-template
4084 subslice of system.slice unless something else is explicitly
4085 configured. For example, instances of sshd@.service will now
4086 implicitly be placed in system-sshd.slice rather than
4087 system.slice as before.
4089 * Test coverage support may now be enabled at build time.
4091 Contributions from: Dave Reisner, Frederic Crozat, Harald
4092 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
4093 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
4094 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
4095 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
4096 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
4097 Giokas, Zbigniew Jędrzejewski-Szmek
4099 — Berlin, 2013-07-23
4103 * Two new unit types have been introduced:
4105 Scope units are very similar to service units, however, are
4106 created out of pre-existing processes — instead of PID 1
4107 forking off the processes. By using scope units it is
4108 possible for system services and applications to group their
4109 own child processes (worker processes) in a powerful way
4110 which then maybe used to organize them, or kill them
4111 together, or apply resource limits on them.
4113 Slice units may be used to partition system resources in an
4114 hierarchical fashion and then assign other units to them. By
4115 default there are now three slices: system.slice (for all
4116 system services), user.slice (for all user sessions),
4117 machine.slice (for VMs and containers).
4119 Slices and scopes have been introduced primarily in
4120 context of the work to move cgroup handling to a
4121 single-writer scheme, where only PID 1
4122 creates/removes/manages cgroups.
4124 * There's a new concept of "transient" units. In contrast to
4125 normal units these units are created via an API at runtime,
4126 not from configuration from disk. More specifically this
4127 means it is now possible to run arbitrary programs as
4128 independent services, with all execution parameters passed
4129 in via bus APIs rather than read from disk. Transient units
4130 make systemd substantially more dynamic then it ever was,
4131 and useful as a general batch manager.
4133 * logind has been updated to make use of scope and slice units
4134 for managing user sessions. As a user logs in he will get
4135 his own private slice unit, to which all sessions are added
4136 as scope units. We also added support for automatically
4137 adding an instance of user@.service for the user into the
4138 slice. Effectively logind will no longer create cgroup
4139 hierarchies on its own now, it will defer entirely to PID 1
4140 for this by means of scope, service and slice units. Since
4141 user sessions this way become entities managed by PID 1
4142 the output of "systemctl" is now a lot more comprehensive.
4144 * A new mini-daemon "systemd-machined" has been added which
4145 may be used by virtualization managers to register local
4146 VMs/containers. nspawn has been updated accordingly, and
4147 libvirt will be updated shortly. machined will collect a bit
4148 of meta information about the VMs/containers, and assign
4149 them their own scope unit (see above). The collected
4150 meta-data is then made available via the "machinectl" tool,
4151 and exposed in "ps" and similar tools. machined/machinectl
4152 is compile-time optional.
4154 * As discussed earlier, the low-level cgroup configuration
4155 options ControlGroup=, ControlGroupModify=,
4156 ControlGroupPersistent=, ControlGroupAttribute= have been
4157 removed. Please use high-level attribute settings instead as
4158 well as slice units.
4160 * A new bus call SetUnitProperties() has been added to alter
4161 various runtime parameters of a unit. This is primarily
4162 useful to alter cgroup parameters dynamically in a nice way,
4163 but will be extended later on to make more properties
4164 modifiable at runtime. systemctl gained a new set-properties
4165 command that wraps this call.
4167 * A new tool "systemd-run" has been added which can be used to
4168 run arbitrary command lines as transient services or scopes,
4169 while configuring a number of settings via the command
4170 line. This tool is currently very basic, however already
4171 very useful. We plan to extend this tool to even allow
4172 queuing of execution jobs with time triggers from the
4173 command line, similar in fashion to "at".
4175 * nspawn will now inform the user explicitly that kernels with
4176 audit enabled break containers, and suggest the user to turn
4179 * Support for detecting the IMA and AppArmor security
4180 frameworks with ConditionSecurity= has been added.
4182 * journalctl gained a new "-k" switch for showing only kernel
4183 messages, mimicking dmesg output; in addition to "--user"
4184 and "--system" switches for showing only user's own logs
4187 * systemd-delta can now show information about drop-in
4188 snippets extending unit files.
4190 * libsystemd-bus has been substantially updated but is still
4191 not available as public API.
4193 * systemd will now look for the "debug" argument on the kernel
4194 command line and enable debug logging, similar to what
4195 "systemd.log_level=debug" already did before.
4197 * "systemctl set-default", "systemctl get-default" has been
4198 added to configure the default.target symlink, which
4199 controls what to boot into by default.
4201 * "systemctl set-log-level" has been added as a convenient
4202 way to raise and lower systemd logging threshold.
4204 * "systemd-analyze plot" will now show the time the various
4205 generators needed for execution, as well as information
4206 about the unit file loading.
4208 * libsystemd-journal gained a new sd_journal_open_files() call
4209 for opening specific journal files. journactl also gained a
4210 new switch to expose this new functionality. Previously we
4211 only supported opening all files from a directory, or all
4212 files from the system, as opening individual files only is
4213 racy due to journal file rotation.
4215 * systemd gained the new DefaultEnvironment= setting in
4216 /etc/systemd/system.conf to set environment variables for
4219 * If a privileged process logs a journal message with the
4220 OBJECT_PID= field set, then journald will automatically
4221 augment this with additional OBJECT_UID=, OBJECT_GID=,
4222 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
4223 system services want to log events about specific client
4224 processes. journactl/systemctl has been updated to make use
4225 of this information if all log messages regarding a specific
4228 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
4229 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
4230 Reisner, David Coppa, David King, David Strauss, Eelco
4231 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
4232 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
4233 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
4234 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
4235 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
4236 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
4237 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
4238 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
4239 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
4240 Łukasz Stelmach, 장동준
4244 * The Python bindings gained some minimal support for the APIs
4245 exposed by libsystemd-logind.
4247 * ConditionSecurity= gained support for detecting SMACK. Since
4248 this condition already supports SELinux and AppArmor we only
4249 miss IMA for this. Patches welcome!
4251 Contributions from: Karol Lewandowski, Lennart Poettering,
4252 Zbigniew Jędrzejewski-Szmek
4256 * systemd-nspawn will now create /etc/resolv.conf if
4257 necessary, before bind-mounting the host's file onto it.
4259 * systemd-nspawn will now store meta information about a
4260 container on the container's cgroup as extended attribute
4261 fields, including the root directory.
4263 * The cgroup hierarchy has been reworked in many ways. All
4264 objects any of the components systemd creates in the cgroup
4265 tree are now suffixed. More specifically, user sessions are
4266 now placed in cgroups suffixed with ".session", users in
4267 cgroups suffixed with ".user", and nspawn containers in
4268 cgroups suffixed with ".nspawn". Furthermore, all cgroup
4269 names are now escaped in a simple scheme to avoid collision
4270 of userspace object names with kernel filenames. This work
4271 is preparation for making these objects relocatable in the
4272 cgroup tree, in order to allow easy resource partitioning of
4273 these objects without causing naming conflicts.
4275 * systemctl list-dependencies gained the new switches
4276 --plain, --reverse, --after and --before.
4278 * systemd-inhibit now shows the process name of processes that
4279 have taken an inhibitor lock.
4281 * nss-myhostname will now also resolve "localhost"
4282 implicitly. This makes /etc/hosts an optional file and
4283 nicely handles that on IPv6 ::1 maps to both "localhost" and
4286 * libsystemd-logind.so gained a new call
4287 sd_get_machine_names() to enumerate running containers and
4288 VMs (currently only supported by very new libvirt and
4289 nspawn). sd_login_monitor can now be used to watch
4290 VMs/containers coming and going.
4292 * .include is not allowed recursively anymore, and only in
4293 unit files. Usually it is better to use drop-in snippets in
4294 .d/*.conf anyway, as introduced with systemd 198.
4296 * systemd-analyze gained a new "critical-chain" command that
4297 determines the slowest chain of units run during system
4298 boot-up. It is very useful for tracking down where
4299 optimizing boot time is the most beneficial.
4301 * systemd will no longer allow manipulating service paths in
4302 the name=systemd:/system cgroup tree using ControlGroup= in
4303 units. (But is still fine with it in all other dirs.)
4305 * There's a new systemd-nspawn@.service service file that may
4306 be used to easily run nspawn containers as system
4307 services. With the container's root directory in
4308 /var/lib/container/foobar it is now sufficient to run
4309 "systemctl start systemd-nspawn@foobar.service" to boot it.
4311 * systemd-cgls gained a new parameter "--machine" to list only
4312 the processes within a certain container.
4314 * ConditionSecurity= now can check for "apparmor". We still
4315 are lacking checks for SMACK and IMA for this condition
4316 check though. Patches welcome!
4318 * A new configuration file /etc/systemd/sleep.conf has been
4319 added that may be used to configure which kernel operation
4320 systemd is supposed to execute when "suspend", "hibernate"
4321 or "hybrid-sleep" is requested. This makes the new kernel
4322 "freeze" state accessible to the user.
4324 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
4325 the passed argument if applicable.
4327 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
4328 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
4329 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
4330 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
4331 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
4332 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
4333 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
4338 * The output of 'systemctl list-jobs' got some polishing. The
4339 '--type=' argument may now be passed more than once. A new
4340 command 'systemctl list-sockets' has been added which shows
4341 a list of kernel sockets systemd is listening on with the
4342 socket units they belong to, plus the units these socket
4345 * The experimental libsystemd-bus library got substantial
4346 updates to work in conjunction with the (also experimental)
4347 kdbus kernel project. It works well enough to exchange
4348 messages with some sophistication. Note that kdbus is not
4349 ready yet, and the library is mostly an elaborate test case
4350 for now, and not installable.
4352 * systemd gained a new unit 'systemd-static-nodes.service'
4353 that generates static device nodes earlier during boot, and
4354 can run in conjunction with udev.
4356 * libsystemd-login gained a new call sd_pid_get_user_unit()
4357 to retrieve the user systemd unit a process is running
4358 in. This is useful for systems where systemd is used as
4361 * systemd-nspawn now places all containers in the new /machine
4362 top-level cgroup directory in the name=systemd
4363 hierarchy. libvirt will soon do the same, so that we get a
4364 uniform separation of /system, /user and /machine for system
4365 services, user processes and containers/virtual
4366 machines. This new cgroup hierarchy is also useful to stick
4367 stable names to specific container instances, which can be
4368 recognized later this way (this name may be controlled
4369 via systemd-nspawn's new -M switch). libsystemd-login also
4370 gained a new call sd_pid_get_machine_name() to retrieve the
4371 name of the container/VM a specific process belongs to.
4373 * bootchart can now store its data in the journal.
4375 * libsystemd-journal gained a new call
4376 sd_journal_add_conjunction() for AND expressions to the
4377 matching logic. This can be used to express more complex
4378 logical expressions.
4380 * journactl can now take multiple --unit= and --user-unit=
4383 * The cryptsetup logic now understands the "luks.key=" kernel
4384 command line switch for specifying a file to read the
4385 decryption key from. Also, if a configured key file is not
4386 found the tool will now automatically fall back to prompting
4389 * Python systemd.journal module was updated to wrap recently
4390 added functions from libsystemd-journal. The interface was
4391 changed to bring the low level interface in s.j._Reader
4392 closer to the C API, and the high level interface in
4393 s.j.Reader was updated to wrap and convert all data about
4396 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
4397 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
4398 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
4399 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
4400 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
4401 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
4405 * journalctl --update-catalog now understands a new --root=
4406 option to operate on catalogs found in a different root
4409 * During shutdown after systemd has terminated all running
4410 services a final killing loop kills all remaining left-over
4411 processes. We will now print the name of these processes
4412 when we send SIGKILL to them, since this usually indicates a
4415 * If /etc/crypttab refers to password files stored on
4416 configured mount points automatic dependencies will now be
4417 generated to ensure the specific mount is established first
4418 before the key file is attempted to be read.
4420 * 'systemctl status' will now show information about the
4421 network sockets a socket unit is listening on.
4423 * 'systemctl status' will also shown information about any
4424 drop-in configuration file for units. (Drop-In configuration
4425 files in this context are files such as
4426 /etc/systemd/systemd/foobar.service.d/*.conf)
4428 * systemd-cgtop now optionally shows summed up CPU times of
4429 cgroups. Press '%' while running cgtop to switch between
4430 percentage and absolute mode. This is useful to determine
4431 which cgroups use up the most CPU time over the entire
4432 runtime of the system. systemd-cgtop has also been updated
4433 to be 'pipeable' for processing with further shell tools.
4435 * 'hostnamectl set-hostname' will now allow setting of FQDN
4438 * The formatting and parsing of time span values has been
4439 changed. The parser now understands fractional expressions
4440 such as "5.5h". The formatter will now output fractional
4441 expressions for all time spans under 1min, i.e. "5.123456s"
4442 rather than "5s 123ms 456us". For time spans under 1s
4443 millisecond values are shown, for those under 1ms
4444 microsecond values are shown. This should greatly improve
4445 all time-related output of systemd.
4447 * libsystemd-login and libsystemd-journal gained new
4448 functions for querying the poll() events mask and poll()
4449 timeout value for integration into arbitrary event
4452 * localectl gained the ability to list available X11 keymaps
4453 (models, layouts, variants, options).
4455 * 'systemd-analyze dot' gained the ability to filter for
4456 specific units via shell-style globs, to create smaller,
4457 more useful graphs. I.e. it is now possible to create simple
4458 graphs of all the dependencies between only target units, or
4459 of all units that Avahi has dependencies with.
4461 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
4462 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
4463 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
4464 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
4465 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
4466 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
4467 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
4471 * The boot-time readahead implementation for rotating media
4472 will now read the read-ahead data in multiple passes which
4473 consist of all read requests made in equidistant time
4474 intervals. This means instead of strictly reading read-ahead
4475 data in its physical order on disk we now try to find a
4476 middle ground between physical and access time order.
4478 * /etc/os-release files gained a new BUILD_ID= field for usage
4479 on operating systems that provide continuous builds of OS
4482 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
4483 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
4484 William Douglas, Zbigniew Jędrzejewski-Szmek
4488 * systemd-python gained an API exposing libsystemd-daemon.
4490 * The SMACK setup logic gained support for uploading CIPSO
4493 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
4494 ReadOnlyDirectories= and InaccessibleDirectories= has
4495 changed. The private /tmp and /var/tmp directories are now
4496 shared by all processes of a service (which means
4497 ExecStartPre= may now leave data in /tmp that ExecStart= of
4498 the same service can still access). When a service is
4499 stopped its temporary directories are immediately deleted
4500 (normal clean-up with tmpfiles is still done in addition to
4503 * By default, systemd will now set a couple of sysctl
4504 variables in the kernel: the safe sysrq options are turned
4505 on, IP route verification is turned on, and source routing
4506 disabled. The recently added hardlink and softlink
4507 protection of the kernel is turned on. These settings should
4508 be reasonably safe, and good defaults for all new systems.
4510 * The predictable network naming logic may now be turned off
4511 with a new kernel command line switch: net.ifnames=0.
4513 * A new libsystemd-bus module has been added that implements a
4514 pretty complete D-Bus client library. For details see:
4516 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
4518 * journald will now explicitly flush the journal files to disk
4519 at the latest 5min after each write. The file will then also
4520 be marked offline until the next write. This should increase
4521 reliability in case of a crash. The synchronization delay
4522 can be configured via SyncIntervalSec= in journald.conf.
4524 * There's a new remote-fs-setup.target unit that can be used
4525 to pull in specific services when at least one remote file
4526 system is to be mounted.
4528 * There are new targets timers.target and paths.target as
4529 canonical targets to pull user timer and path units in
4530 from. This complements sockets.target with a similar
4531 purpose for socket units.
4533 * libudev gained a new call udev_device_set_attribute_value()
4534 to set sysfs attributes of a device.
4536 * The udev daemon now sets the default number of worker
4537 processes executed in parallel based on the number of available
4538 CPUs instead of the amount of available RAM. This is supposed
4539 to provide a more reliable default and limit a too aggressive
4540 paralellism for setups with 1000s of devices connected.
4542 Contributions from: Auke Kok, Colin Walters, Cristian
4543 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
4544 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
4545 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
4546 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
4547 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
4548 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
4549 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
4550 Zbigniew Jędrzejewski-Szmek
4554 * Configuration of unit files may now be extended via drop-in
4555 files without having to edit/override the unit files
4556 themselves. More specifically, if the administrator wants to
4557 change one value for a service file foobar.service he can
4558 now do so by dropping in a configuration snippet into
4559 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
4560 will load all these snippets and apply them on top of the
4561 main unit configuration file, possibly extending or
4562 overriding its settings. Using these drop-in snippets is
4563 generally nicer than the two earlier options for changing
4564 unit files locally: copying the files from
4565 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
4566 them there; or creating a new file in /etc/systemd/system/
4567 that incorporates the original one via ".include". Drop-in
4568 snippets into these .d/ directories can be placed in any
4569 directory systemd looks for units in, and the usual
4570 overriding semantics between /usr/lib, /etc and /run apply
4573 * Most unit file settings which take lists of items can now be
4574 reset by assigning the empty string to them. For example,
4575 normally, settings such as Environment=FOO=BAR append a new
4576 environment variable assignment to the environment block,
4577 each time they are used. By assigning Environment= the empty
4578 string the environment block can be reset to empty. This is
4579 particularly useful with the .d/*.conf drop-in snippets
4580 mentioned above, since this adds the ability to reset list
4581 settings from vendor unit files via these drop-ins.
4583 * systemctl gained a new "list-dependencies" command for
4584 listing the dependencies of a unit recursively.
4586 * Inhibitors are now honored and listed by "systemctl
4587 suspend", "systemctl poweroff" (and similar) too, not only
4588 GNOME. These commands will also list active sessions by
4591 * Resource limits (as exposed by the various control group
4592 controllers) can now be controlled dynamically at runtime
4593 for all units. More specifically, you can now use a command
4594 like "systemctl set-cgroup-attr foobar.service cpu.shares
4595 2000" to alter the CPU shares a specific service gets. These
4596 settings are stored persistently on disk, and thus allow the
4597 administrator to easily adjust the resource usage of
4598 services with a few simple commands. This dynamic resource
4599 management logic is also available to other programs via the
4600 bus. Almost any kernel cgroup attribute and controller is
4603 * systemd-vconsole-setup will now copy all font settings to
4604 all allocated VTs, where it previously applied them only to
4607 * libsystemd-login gained the new sd_session_get_tty() API
4610 * This release drops support for a few legacy or
4611 distribution-specific LSB facility names when parsing init
4612 scripts: $x-display-manager, $mail-transfer-agent,
4613 $mail-transport-agent, $mail-transfer-agent, $smtp,
4614 $null. Also, the mail-transfer-agent.target unit backing
4615 this has been removed. Distributions which want to retain
4616 compatibility with this should carry the burden for
4617 supporting this themselves and patch support for these back
4618 in, if they really need to. Also, the facilities $syslog and
4619 $local_fs are now ignored, since systemd does not support
4620 early-boot LSB init scripts anymore, and these facilities
4621 are implied anyway for normal services. syslog.target has
4624 * There are new bus calls on PID1's Manager object for
4625 cancelling jobs, and removing snapshot units. Previously,
4626 both calls were only available on the Job and Snapshot
4629 * systemd-journal-gatewayd gained SSL support.
4631 * The various "environment" files, such as /etc/locale.conf
4632 now support continuation lines with a backslash ("\") as
4633 last character in the line, similarly in style (but different)
4634 to how this is supported in shells.
4636 * For normal user processes the _SYSTEMD_USER_UNIT= field is
4637 now implicitly appended to every log entry logged. systemctl
4638 has been updated to filter by this field when operating on a
4639 user systemd instance.
4641 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
4642 CAP_AUDIT_CONTROL capabilities to the capabilities set for
4643 the container. This makes it easier to boot unmodified
4644 Fedora systems in a container, which however still requires
4645 audit=0 to be passed on the kernel command line. Auditing in
4646 kernel and userspace is unfortunately still too broken in
4647 context of containers, hence we recommend compiling it out
4648 of the kernel or using audit=0. Hopefully this will be fixed
4649 one day for good in the kernel.
4651 * nspawn gained the new --bind= and --bind-ro= parameters to
4652 bind mount specific directories from the host into the
4655 * nspawn will now mount its own devpts file system instance
4656 into the container, in order not to leak pty devices from
4657 the host into the container.
4659 * systemd will now read the firmware boot time performance
4660 information from the EFI variables, if the used boot loader
4661 supports this, and takes it into account for boot performance
4662 analysis via "systemd-analyze". This is currently supported
4663 only in conjunction with Gummiboot, but could be supported
4664 by other boot loaders too. For details see:
4666 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
4668 * A new generator has been added that automatically mounts the
4669 EFI System Partition (ESP) to /boot, if that directory
4670 exists, is empty, and no other file system has been
4671 configured to be mounted there.
4673 * logind will now send out PrepareForSleep(false) out
4674 unconditionally, after coming back from suspend. This may be
4675 used by applications as asynchronous notification for
4676 system resume events.
4678 * "systemctl unlock-sessions" has been added, that allows
4679 unlocking the screens of all user sessions at once, similar
4680 to how "systemctl lock-sessions" already locked all users
4681 sessions. This is backed by a new D-Bus call UnlockSessions().
4683 * "loginctl seat-status" will now show the master device of a
4684 seat. (i.e. the device of a seat that needs to be around for
4685 the seat to be considered available, usually the graphics
4688 * tmpfiles gained a new "X" line type, that allows
4689 configuration of files and directories (with wildcards) that
4690 shall be excluded from automatic cleanup ("aging").
4692 * udev default rules set the device node permissions now only
4693 at "add" events, and do not change them any longer with a
4694 later "change" event.
4696 * The log messages for lid events and power/sleep keypresses
4697 now carry a message ID.
4699 * We now have a substantially larger unit test suite, but this
4700 continues to be work in progress.
4702 * udevadm hwdb gained a new --root= parameter to change the
4703 root directory to operate relative to.
4705 * logind will now issue a background sync() request to the kernel
4706 early at shutdown, so that dirty buffers are flushed to disk early
4707 instead of at the last moment, in order to optimize shutdown
4710 * A new bootctl tool has been added that is an interface for
4711 certain boot loader operations. This is currently a preview
4712 and is likely to be extended into a small mechanism daemon
4713 like timedated, localed, hostnamed, and can be used by
4714 graphical UIs to enumerate available boot options, and
4715 request boot into firmware operations.
4717 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
4718 the rest of the package. It also has been updated to work
4719 correctly in initrds.
4721 * Policykit previously has been runtime optional, and is now
4722 also compile time optional via a configure switch.
4724 * systemd-analyze has been reimplemented in C. Also "systemctl
4725 dot" has moved into systemd-analyze.
4727 * "systemctl status" with no further parameters will now print
4728 the status of all active or failed units.
4730 * Operations such as "systemctl start" can now be executed
4731 with a new mode "--irreversible" which may be used to queue
4732 operations that cannot accidentally be reversed by a later
4733 job queuing. This is by default used to make shutdown
4734 requests more robust.
4736 * The Python API of systemd now gained a new module for
4737 reading journal files.
4739 * A new tool kernel-install has been added that can install
4740 kernel images according to the Boot Loader Specification:
4742 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
4744 * Boot time console output has been improved to provide
4745 animated boot time output for hanging jobs.
4747 * A new tool systemd-activate has been added which can be used
4748 to test socket activation with, directly from the command
4749 line. This should make it much easier to test and debug
4750 socket activation in daemons.
4752 * journalctl gained a new "--reverse" (or -r) option to show
4753 journal output in reverse order (i.e. newest line first).
4755 * journalctl gained a new "--pager-end" (or -e) option to jump
4756 to immediately jump to the end of the journal in the
4757 pager. This is only supported in conjunction with "less".
4759 * journalctl gained a new "--user-unit=" option, that works
4760 similarly to "--unit=" but filters for user units rather than
4763 * A number of unit files to ease adoption of systemd in
4764 initrds has been added. This moves some minimal logic from
4765 the various initrd implementations into systemd proper.
4767 * The journal files are now owned by a new group
4768 "systemd-journal", which exists specifically to allow access
4769 to the journal, and nothing else. Previously, we used the
4770 "adm" group for that, which however possibly covers more
4771 than just journal/log file access. This new group is now
4772 already used by systemd-journal-gatewayd to ensure this
4773 daemon gets access to the journal files and as little else
4774 as possible. Note that "make install" will also set FS ACLs
4775 up for /var/log/journal to give "adm" and "wheel" read
4776 access to it, in addition to "systemd-journal" which owns
4777 the journal files. We recommend that packaging scripts also
4778 add read access to "adm" + "wheel" to /var/log/journal, and
4779 all existing/future journal files. To normal users and
4780 administrators little changes, however packagers need to
4781 ensure to create the "systemd-journal" system group at
4782 package installation time.
4784 * The systemd-journal-gatewayd now runs as unprivileged user
4785 systemd-journal-gateway:systemd-journal-gateway. Packaging
4786 scripts need to create these system user/group at
4789 * timedated now exposes a new boolean property CanNTP that
4790 indicates whether a local NTP service is available or not.
4792 * systemd-detect-virt will now also detect xen PVs
4794 * The pstore file system is now mounted by default, if it is
4797 * In addition to the SELinux and IMA policies we will now also
4798 load SMACK policies at early boot.
4800 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
4801 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
4802 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
4803 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
4804 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
4805 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
4806 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
4807 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
4808 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
4809 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
4810 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
4811 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
4812 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
4813 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
4817 * Timer units now support calendar time events in addition to
4818 monotonic time events. That means you can now trigger a unit
4819 based on a calendar time specification such as "Thu,Fri
4820 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
4821 or fifth day of any month of the year 2013, given that it is
4822 a thursday or friday. This brings timer event support
4823 considerably closer to cron's capabilities. For details on
4824 the supported calendar time specification language see
4827 * udev now supports a number of different naming policies for
4828 network interfaces for predictable names, and a combination
4829 of these policies is now the default. Please see this wiki
4830 document for details:
4832 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
4834 * Auke Kok's bootchart implementation has been added to the
4835 systemd tree. It is an optional component that can graph the
4836 boot in quite some detail. It is one of the best bootchart
4837 implementations around and minimal in its code and
4840 * nss-myhostname has been integrated into the systemd source
4841 tree. nss-myhostname guarantees that the local hostname
4842 always stays resolvable via NSS. It has been a weak
4843 requirement of systemd-hostnamed since a long time, and
4844 since its code is actually trivial we decided to just
4845 include it in systemd's source tree. It can be turned off
4846 with a configure switch.
4848 * The read-ahead logic is now capable of properly detecting
4849 whether a btrfs file system is on SSD or rotating media, in
4850 order to optimize the read-ahead scheme. Previously, it was
4851 only capable of detecting this on traditional file systems
4854 * In udev, additional device properties are now read from the
4855 IAB in addition to the OUI database. Also, Bluetooth company
4856 identities are attached to the devices as well.
4858 * In service files %U may be used as specifier that is
4859 replaced by the configured user name of the service.
4861 * nspawn may now be invoked without a controlling TTY. This
4862 makes it suitable for invocation as its own service. This
4863 may be used to set up a simple containerized server system
4864 using only core OS tools.
4866 * systemd and nspawn can now accept socket file descriptors
4867 when they are started for socket activation. This enables
4868 implementation of socket activated nspawn
4869 containers. i.e. think about autospawning an entire OS image
4870 when the first SSH or HTTP connection is received. We expect
4871 that similar functionality will also be added to libvirt-lxc
4874 * journalctl will now suppress ANSI color codes when
4875 presenting log data.
4877 * systemctl will no longer show control group information for
4878 a unit if a the control group is empty anyway.
4880 * logind can now automatically suspend/hibernate/shutdown the
4883 * /etc/machine-info and hostnamed now also expose the chassis
4884 type of the system. This can be used to determine whether
4885 the local system is a laptop, desktop, handset or
4886 tablet. This information may either be configured by the
4887 user/vendor or is automatically determined from ACPI and DMI
4888 information if possible.
4890 * A number of PolicyKit actions are now bound together with
4891 "imply" rules. This should simplify creating UIs because
4892 many actions will now authenticate similar ones as well.
4894 * Unit files learnt a new condition ConditionACPower= which
4895 may be used to conditionalize a unit depending on whether an
4896 AC power source is connected or not, of whether the system
4897 is running on battery power.
4899 * systemctl gained a new "is-failed" verb that may be used in
4900 shell scripts and suchlike to check whether a specific unit
4901 is in the "failed" state.
4903 * The EnvironmentFile= setting in unit files now supports file
4904 globbing, and can hence be used to easily read a number of
4905 environment files at once.
4907 * systemd will no longer detect and recognize specific
4908 distributions. All distribution-specific #ifdeffery has been
4909 removed, systemd is now fully generic and
4910 distribution-agnostic. Effectively, not too much is lost as
4911 a lot of the code is still accessible via explicit configure
4912 switches. However, support for some distribution specific
4913 legacy configuration file formats has been dropped. We
4914 recommend distributions to simply adopt the configuration
4915 files everybody else uses now and convert the old
4916 configuration from packaging scripts. Most distributions
4917 already did that. If that's not possible or desirable,
4918 distributions are welcome to forward port the specific
4919 pieces of code locally from the git history.
4921 * When logging a message about a unit systemd will now always
4922 log the unit name in the message meta data.
4924 * localectl will now also discover system locale data that is
4925 not stored in locale archives, but directly unpacked.
4927 * logind will no longer unconditionally use framebuffer
4928 devices as seat masters, i.e. as devices that are required
4929 to be existing before a seat is considered preset. Instead,
4930 it will now look for all devices that are tagged as
4931 "seat-master" in udev. By default, framebuffer devices will
4932 be marked as such, but depending on local systems, other
4933 devices might be marked as well. This may be used to
4934 integrate graphics cards using closed source drivers (such
4935 as NVidia ones) more nicely into logind. Note however, that
4936 we recommend using the open source NVidia drivers instead,
4937 and no udev rules for the closed-source drivers will be
4938 shipped from us upstream.
4940 Contributions from: Adam Williamson, Alessandro Crismani, Auke
4941 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
4942 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
4943 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
4944 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
4945 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
4946 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
4947 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
4948 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
4949 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
4950 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
4951 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
4956 * udev gained support for loading additional device properties
4957 from an indexed database that is keyed by vendor/product IDs
4958 and similar device identifiers. For the beginning this
4959 "hwdb" is populated with data from the well-known PCI and
4960 USB database, but also includes PNP, ACPI and OID data. In
4961 the longer run this indexed database shall grow into
4962 becoming the one central database for non-essential
4963 userspace device metadata. Previously, data from the PCI/USB
4964 database was only attached to select devices, since the
4965 lookup was a relatively expensive operation due to O(n) time
4966 complexity (with n being the number of entries in the
4967 database). Since this is now O(1), we decided to add in this
4968 data for all devices where this is available, by
4969 default. Note that the indexed database needs to be rebuilt
4970 when new data files are installed. To achieve this you need
4971 to update your packaging scripts to invoke "udevadm hwdb
4972 --update" after installation of hwdb data files. For
4973 RPM-based distributions we introduced the new
4974 %udev_hwdb_update macro for this purpose.
4976 * The Journal gained support for the "Message Catalog", an
4977 indexed database to link up additional information with
4978 journal entries. For further details please check:
4980 http://www.freedesktop.org/wiki/Software/systemd/catalog
4982 The indexed message catalog database also needs to be
4983 rebuilt after installation of message catalog files. Use
4984 "journalctl --update-catalog" for this. For RPM-based
4985 distributions we introduced the %journal_catalog_update
4986 macro for this purpose.
4988 * The Python Journal bindings gained support for the standard
4989 Python logging framework.
4991 * The Journal API gained new functions for checking whether
4992 the underlying file system of a journal file is capable of
4993 properly reporting file change notifications, or whether
4994 applications that want to reflect journal changes "live"
4995 need to recheck journal files continuously in appropriate
4998 * It is now possible to set the "age" field for tmpfiles
4999 entries to 0, indicating that files matching this entry
5000 shall always be removed when the directories are cleaned up.
5002 * coredumpctl gained a new "gdb" verb which invokes gdb
5003 right-away on the selected coredump.
5005 * There's now support for "hybrid sleep" on kernels that
5006 support this, in addition to "suspend" and "hibernate". Use
5007 "systemctl hybrid-sleep" to make use of this.
5009 * logind's HandleSuspendKey= setting (and related settings)
5010 now gained support for a new "lock" setting to simply
5011 request the screen lock on all local sessions, instead of
5012 actually executing a suspend or hibernation.
5014 * systemd will now mount the EFI variables file system by
5017 * Socket units now gained support for configuration of the
5018 SMACK security label.
5020 * timedatectl will now output the time of the last and next
5021 daylight saving change.
5023 * We dropped support for various legacy and distro-specific
5024 concepts, such as insserv, early-boot SysV services
5025 (i.e. those for non-standard runlevels such as 'b' or 'S')
5026 or ArchLinux /etc/rc.conf support. We recommend the
5027 distributions who still need support this to either continue
5028 to maintain the necessary patches downstream, or find a
5029 different solution. (Talk to us if you have questions!)
5031 * Various systemd components will now bypass PolicyKit checks
5032 for root and otherwise handle properly if PolicyKit is not
5033 found to be around. This should fix most issues for
5034 PolicyKit-less systems. Quite frankly this should have been
5035 this way since day one. It is absolutely our intention to
5036 make systemd work fine on PolicyKit-less systems, and we
5037 consider it a bug if something does not work as it should if
5038 PolicyKit is not around.
5040 * For embedded systems it is now possible to build udev and
5041 systemd without blkid and/or kmod support.
5043 * "systemctl switch-root" is now capable of switching root
5044 more than once. I.e. in addition to transitions from the
5045 initrd to the host OS it is now possible to transition to
5046 further OS images from the host. This is useful to implement
5047 offline updating tools.
5049 * Various other additions have been made to the RPM macros
5050 shipped with systemd. Use %udev_rules_update() after
5051 installing new udev rules files. %_udevhwdbdir,
5052 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
5053 %_sysctldir are now available which resolve to the right
5054 directories for packages to place various data files in.
5056 * journalctl gained the new --full switch (in addition to
5057 --all, to disable ellipsation for long messages.
5059 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
5060 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
5061 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
5062 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
5063 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
5064 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
5065 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
5066 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
5067 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
5071 * journalctl gained new --since= and --until= switches to
5072 filter by time. It also now supports nice filtering for
5073 units via --unit=/-u.
5075 * Type=oneshot services may use ExecReload= and do the
5078 * The journal daemon now supports time-based rotation and
5079 vacuuming, in addition to the usual disk-space based
5082 * The journal will now index the available field values for
5083 each field name. This enables clients to show pretty drop
5084 downs of available match values when filtering. The bash
5085 completion of journalctl has been updated
5086 accordingly. journalctl gained a new switch -F to list all
5087 values a certain field takes in the journal database.
5089 * More service events are now written as structured messages
5090 to the journal, and made recognizable via message IDs.
5092 * The timedated, localed and hostnamed mini-services which
5093 previously only provided support for changing time, locale
5094 and hostname settings from graphical DEs such as GNOME now
5095 also have a minimal (but very useful) text-based client
5096 utility each. This is probably the nicest way to changing
5097 these settings from the command line now, especially since
5098 it lists available options and is fully integrated with bash
5101 * There's now a new tool "systemd-coredumpctl" to list and
5102 extract coredumps from the journal.
5104 * We now install a README each in /var/log/ and
5105 /etc/rc.d/init.d explaining where the system logs and init
5106 scripts went. This hopefully should help folks who go to
5107 that dirs and look into the otherwise now empty void and
5108 scratch their heads.
5110 * When user-services are invoked (by systemd --user) the
5111 $MANAGERPID env var is set to the PID of systemd.
5113 * SIGRTMIN+24 when sent to a --user instance will now result
5114 in immediate termination of systemd.
5116 * gatewayd received numerous feature additions such as a
5117 "follow" mode, for live syncing and filtering.
5119 * browse.html now allows filtering and showing detailed
5120 information on specific entries. Keyboard navigation and
5121 mouse screen support has been added.
5123 * gatewayd/journalctl now supports HTML5/JSON
5124 Server-Sent-Events as output.
5126 * The SysV init script compatibility logic will now
5127 heuristically determine whether a script supports the
5128 "reload" verb, and only then make this available as
5131 * "systemctl status --follow" has been removed, use "journalctl
5134 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
5135 have been removed since they are hardly useful to be
5138 * And I'd like to take the opportunity to specifically mention
5139 Zbigniew for his great contributions. Zbigniew, you rock!
5141 Contributions from: Andrew Eikum, Christian Hesse, Colin
5142 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
5143 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
5144 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
5145 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
5146 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
5147 Jędrzejewski-Szmek, Сковорода Никита Андреевич
5151 * If /etc/vconsole.conf is non-existent or empty we will no
5152 longer load any console font or key map at boot by
5153 default. Instead the kernel defaults will be left
5154 intact. This is definitely the right thing to do, as no
5155 configuration should mean no configuration, and hard-coding
5156 font names that are different on all archs is probably a bad
5157 idea. Also, the kernel default key map and font should be
5158 good enough for most cases anyway, and mostly identical to
5159 the userspace fonts/key maps we previously overloaded them
5160 with. If distributions want to continue to default to a
5161 non-kernel font or key map they should ship a default
5162 /etc/vconsole.conf with the appropriate contents.
5164 Contributions from: Colin Walters, Daniel J Walsh, Dave
5165 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
5166 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
5170 * journalctl gained a new --cursor= switch to show entries
5171 starting from the specified location in the journal.
5173 * We now enforce a size limit on journal entry fields exported
5174 with "-o json" in journalctl. Fields larger than 4K will be
5175 assigned null. This can be turned off with --all.
5177 * An (optional) journal gateway daemon is now available as
5178 "systemd-journal-gatewayd.service". This service provides
5179 access to the journal via HTTP and JSON. This functionality
5180 will be used to implement live log synchronization in both
5181 pull and push modes, but has various other users too, such
5182 as easy log access for debugging of embedded devices. Right
5183 now it is already useful to retrieve the journal via HTTP:
5185 # systemctl start systemd-journal-gatewayd.service
5186 # wget http://localhost:19531/entries
5188 This will download the journal contents in a
5189 /var/log/messages compatible format. The same as JSON:
5191 # curl -H"Accept: application/json" http://localhost:19531/entries
5193 This service is also accessible via a web browser where a
5194 single static HTML5 app is served that uses the JSON logic
5195 to enable the user to do some basic browsing of the
5196 journal. This will be extended later on. Here's an example
5197 screenshot of this app in its current state:
5199 http://0pointer.de/public/journal-gatewayd
5201 Contributions from: Kay Sievers, Lennart Poettering, Robert
5202 Milasan, Tom Gundersen
5206 * The bash completion logic is now available for journalctl
5209 * We do not mount the "cpuset" controller anymore together with
5210 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
5211 started if no parameters are assigned to it. "cpuset" hence
5212 broke code that assumed it it could create "cpu" groups and
5215 * journalctl -f will now subscribe to terminal size changes,
5216 and line break accordingly.
5218 Contributions from: Dave Reisner, Kay Sievers, Lennart
5219 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
5223 * nspawn will now create a symlink /etc/localtime in the
5224 container environment, copying the host's timezone
5225 setting. Previously this has been done via a bind mount, but
5226 since symlinks cannot be bind mounted this has now been
5227 changed to create/update the appropriate symlink.
5229 * journalctl -n's line number argument is now optional, and
5230 will default to 10 if omitted.
5232 * journald will now log the maximum size the journal files may
5233 take up on disk. This is particularly useful if the default
5234 built-in logic of determining this parameter from the file
5235 system size is used. Use "systemctl status
5236 systemd-journald.service" to see this information.
5238 * The multi-seat X wrapper tool has been stripped down. As X
5239 is now capable of enumerating graphics devices via udev in a
5240 seat-aware way the wrapper is not strictly necessary
5241 anymore. A stripped down temporary stop-gap is still shipped
5242 until the upstream display managers have been updated to
5243 fully support the new X logic. Expect this wrapper to be
5244 removed entirely in one of the next releases.
5246 * HandleSleepKey= in logind.conf has been split up into
5247 HandleSuspendKey= and HandleHibernateKey=. The old setting
5248 is not available anymore. X11 and the kernel are
5249 distinguishing between these keys and we should too. This
5250 also means the inhibition lock for these keys has been split
5253 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
5254 Poettering, Lukas Nykryn, Václav Pavlín
5258 * Whenever a unit changes state we will now log this to the
5259 journal and show along the unit's own log output in
5262 * ConditionPathIsMountPoint= can now properly detect bind
5263 mount points too. (Previously, a bind mount of one file
5264 system to another place in the same file system could not be
5265 detected as mount, since they shared struct stat's st_dev
5268 * We will now mount the cgroup controllers cpu, cpuacct,
5269 cpuset and the controllers net_cls, net_prio together by
5272 * nspawn containers will now have a virtualized boot
5273 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
5274 over with a randomized ID at container initialization). This
5275 has the effect of making "journalctl -b" do the right thing
5278 * The JSON output journal serialization has been updated not
5279 to generate "endless" list objects anymore, but rather one
5280 JSON object per line. This is more in line how most JSON
5281 parsers expect JSON objects. The new output mode
5282 "json-pretty" has been added to provide similar output, but
5283 neatly aligned for readability by humans.
5285 * We dropped all explicit sync() invocations in the shutdown
5286 code. The kernel does this implicitly anyway in the kernel
5287 reboot() syscall. halt(8)'s -n option is now a compatibility
5290 * We now support virtualized reboot() in containers, as
5291 supported by newer kernels. We will fall back to exit() if
5292 CAP_SYS_REBOOT is not available to the container. Also,
5293 nspawn makes use of this now and will actually reboot the
5294 container if the containerized OS asks for that.
5296 * journalctl will only show local log output by default
5297 now. Use --merge (-m) to show remote log output, too.
5299 * libsystemd-journal gained the new sd_journal_get_usage()
5300 call to determine the current disk usage of all journal
5301 files. This is exposed in the new "journalctl --disk-usage"
5304 * journald gained a new configuration setting SplitMode= in
5305 journald.conf which may be used to control how user journals
5306 are split off. See journald.conf(5) for details.
5308 * A new condition type ConditionFileNotEmpty= has been added.
5310 * tmpfiles' "w" lines now support file globbing, to write
5311 multiple files at once.
5313 * We added Python bindings for the journal submission
5314 APIs. More Python APIs for a number of selected APIs will
5315 likely follow. Note that we intend to add native bindings
5316 only for the Python language, as we consider it common
5317 enough to deserve bindings shipped within systemd. There are
5318 various projects outside of systemd that provide bindings
5319 for languages such as PHP or Lua.
5321 * Many conditions will now resolve specifiers such as %i. In
5322 addition, PathChanged= and related directives of .path units
5323 now support specifiers as well.
5325 * There's now a new RPM macro definition for the system preset
5328 * journald will now warn if it ca not forward a message to the
5329 syslog daemon because its socket is full.
5331 * timedated will no longer write or process /etc/timezone,
5332 except on Debian. As we do not support late mounted /usr
5333 anymore /etc/localtime always being a symlink is now safe,
5334 and hence the information in /etc/timezone is not necessary
5337 * logind will now always reserve one VT for a text getty (VT6
5338 by default). Previously if more than 6 X sessions where
5339 started they took up all the VTs with auto-spawned gettys,
5340 so that no text gettys were available anymore.
5342 * udev will now automatically inform the btrfs kernel logic
5343 about btrfs RAID components showing up. This should make
5344 simple hotplug based btrfs RAID assembly work.
5346 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
5347 (but not for its children which will stay at the kernel
5348 default). This should allow setups with a lot more listening
5351 * systemd will now always pass the configured timezone to the
5352 kernel at boot. timedated will do the same when the timezone
5355 * logind's inhibition logic has been updated. By default,
5356 logind will now handle the lid switch, the power and sleep
5357 keys all the time, even in graphical sessions. If DEs want
5358 to handle these events on their own they should take the new
5359 handle-power-key, handle-sleep-key and handle-lid-switch
5360 inhibitors during their runtime. A simple way to achieve
5361 that is to invoke the DE wrapped in an invocation of:
5363 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
5365 * Access to unit operations is now checked via SELinux taking
5366 the unit file label and client process label into account.
5368 * systemd will now notify the administrator in the journal
5369 when he over-mounts a non-empty directory.
5371 * There are new specifiers that are resolved in unit files,
5372 for the host name (%H), the machine ID (%m) and the boot ID
5375 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
5376 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
5377 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
5378 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
5379 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
5380 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5381 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
5385 * Support for reading structured kernel messages from
5386 /dev/kmsg has now been added and is enabled by default.
5388 * Support for reading kernel messages from /proc/kmsg has now
5389 been removed. If you want kernel messages in the journal
5390 make sure to run a recent kernel (>= 3.5) that supports
5391 reading structured messages from /dev/kmsg (see
5392 above). /proc/kmsg is now exclusive property of classic
5393 syslog daemons again.
5395 * The libudev API gained the new
5396 udev_device_new_from_device_id() call.
5398 * The logic for file system namespace (ReadOnlyDirectory=,
5399 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
5400 require pivot_root() anymore. This means fewer temporary
5401 directories are created below /tmp for this feature.
5403 * nspawn containers will now see and receive all submounts
5404 made on the host OS below the root file system of the
5407 * Forward Secure Sealing is now supported for Journal files,
5408 which provide cryptographical sealing of journal files so
5409 that attackers cannot alter log history anymore without this
5410 being detectable. Lennart will soon post a blog story about
5411 this explaining it in more detail.
5413 * There are two new service settings RestartPreventExitStatus=
5414 and SuccessExitStatus= which allow configuration of exit
5415 status (exit code or signal) which will be excepted from the
5416 restart logic, resp. consider successful.
5418 * journalctl gained the new --verify switch that can be used
5419 to check the integrity of the structure of journal files and
5420 (if Forward Secure Sealing is enabled) the contents of
5423 * nspawn containers will now be run with /dev/stdin, /dev/fd/
5424 and similar symlinks pre-created. This makes running shells
5425 as container init process a lot more fun.
5427 * The fstab support can now handle PARTUUID= and PARTLABEL=
5430 * A new ConditionHost= condition has been added to match
5431 against the hostname (with globs) and machine ID. This is
5432 useful for clusters where a single OS image is used to
5433 provision a large number of hosts which shall run slightly
5434 different sets of services.
5436 * Services which hit the restart limit will now be placed in a
5439 Contributions from: Bertram Poettering, Dave Reisner, Huang
5440 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
5441 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
5445 * When running in --user mode systemd will now become a
5446 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
5447 tree a lot more organized.
5449 * A new PartOf= unit dependency type has been introduced that
5450 may be used to group services in a natural way.
5452 * "systemctl enable" may now be used to enable instances of
5455 * journalctl now prints error log levels in red, and
5456 warning/notice log levels in bright white. It also supports
5457 filtering by log level now.
5459 * cgtop gained a new -n switch (similar to top), to configure
5460 the maximum number of iterations to run for. It also gained
5461 -b, to run in batch mode (accepting no input).
5463 * The suffix ".service" may now be omitted on most systemctl
5464 command lines involving service unit names.
5466 * There's a new bus call in logind to lock all sessions, as
5467 well as a loginctl verb for it "lock-sessions".
5469 * libsystemd-logind.so gained a new call sd_journal_perror()
5470 that works similar to libc perror() but logs to the journal
5471 and encodes structured information about the error number.
5473 * /etc/crypttab entries now understand the new keyfile-size=
5476 * shutdown(8) now can send a (configurable) wall message when
5477 a shutdown is cancelled.
5479 * The mount propagation mode for the root file system will now
5480 default to "shared", which is useful to make containers work
5481 nicely out-of-the-box so that they receive new mounts from
5482 the host. This can be undone locally by running "mount
5483 --make-rprivate /" if needed.
5485 * The prefdm.service file has been removed. Distributions
5486 should maintain this unit downstream if they intend to keep
5487 it around. However, we recommend writing normal unit files
5488 for display managers instead.
5490 * Since systemd is a crucial part of the OS we will now
5491 default to a number of compiler switches that improve
5492 security (hardening) such as read-only relocations, stack
5493 protection, and suchlike.
5495 * The TimeoutSec= setting for services is now split into
5496 TimeoutStartSec= and TimeoutStopSec= to allow configuration
5497 of individual time outs for the start and the stop phase of
5500 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
5501 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
5502 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
5503 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
5504 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
5505 Gundersen, Zbigniew Jędrzejewski-Szmek
5509 * The journal and id128 C APIs are now fully documented as man
5512 * Extra safety checks have been added when transitioning from
5513 the initial RAM disk to the main system to avoid accidental
5516 * /etc/crypttab entries now understand the new keyfile-offset=
5519 * systemctl -t can now be used to filter by unit load state.
5521 * The journal C API gained the new sd_journal_wait() call to
5522 make writing synchronous journal clients easier.
5524 * journalctl gained the new -D switch to show journals from a
5527 * journalctl now displays a special marker between log
5528 messages of two different boots.
5530 * The journal is now explicitly flushed to /var via a service
5531 systemd-journal-flush.service, rather than implicitly simply
5532 by seeing /var/log/journal to be writable.
5534 * journalctl (and the journal C APIs) can now match for much
5535 more complex expressions, with alternatives and
5538 * When transitioning from the initial RAM disk to the main
5539 system we will now kill all processes in a killing spree to
5540 ensure no processes stay around by accident.
5542 * Three new specifiers may be used in unit files: %u, %h, %s
5543 resolve to the user name, user home directory resp. user
5544 shell. This is useful for running systemd user instances.
5546 * We now automatically rotate journal files if their data
5547 object hash table gets a fill level > 75%. We also size the
5548 hash table based on the configured maximum file size. This
5549 together should lower hash collisions drastically and thus
5550 speed things up a bit.
5552 * journalctl gained the new "--header" switch to introspect
5553 header data of journal files.
5555 * A new setting SystemCallFilters= has been added to services
5556 which may be used to apply blacklists or whitelists to
5557 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
5559 * nspawn gained a new --link-journal= switch (and quicker: -j)
5560 to link the container journal with the host. This makes it
5561 very easy to centralize log viewing on the host for all
5562 guests while still keeping the journal files separated.
5564 * Many bugfixes and optimizations
5566 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
5567 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
5568 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
5573 * Several tools now understand kernel command line arguments,
5574 which are only read when run in an initial RAM disk. They
5575 usually follow closely their normal counterparts, but are
5578 * There's a new tool to analyze the readahead files that are
5579 automatically generated at boot. Use:
5581 /usr/lib/systemd/systemd-readahead analyze /.readahead
5583 * We now provide an early debug shell on tty9 if this enabled. Use:
5585 systemctl enable debug-shell.service
5587 * All plymouth related units have been moved into the Plymouth
5588 package. Please make sure to upgrade your Plymouth version
5591 * systemd-tmpfiles now supports getting passed the basename of
5592 a configuration file only, in which case it will look for it
5593 in all appropriate directories automatically.
5595 * udevadm info now takes a /dev or /sys path as argument, and
5596 does the right thing. Example:
5598 udevadm info /dev/sda
5599 udevadm info /sys/class/block/sda
5601 * systemctl now prints a warning if a unit is stopped but a
5602 unit that might trigger it continues to run. Example: a
5603 service is stopped but the socket that activates it is left
5606 * "systemctl status" will now mention if the log output was
5607 shortened due to rotation since a service has been started.
5609 * The journal API now exposes functions to determine the
5610 "cutoff" times due to rotation.
5612 * journald now understands SIGUSR1 and SIGUSR2 for triggering
5613 immediately flushing of runtime logs to /var if possible,
5614 resp. for triggering immediate rotation of the journal
5617 * It is now considered an error if a service is attempted to
5618 be stopped that is not loaded.
5620 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
5622 * systemd-analyze now supports Python 3
5624 * tmpfiles now supports cleaning up directories via aging
5625 where the first level dirs are always kept around but
5626 directories beneath it automatically aged. This is enabled
5627 by prefixing the age field with '~'.
5629 * Seat objects now expose CanGraphical, CanTTY properties
5630 which is required to deal with very fast bootups where the
5631 display manager might be running before the graphics drivers
5632 completed initialization.
5634 * Seat objects now expose a State property.
5636 * We now include RPM macros for service enabling/disabling
5637 based on the preset logic. We recommend RPM based
5638 distributions to make use of these macros if possible. This
5639 makes it simpler to reuse RPM spec files across
5642 * We now make sure that the collected systemd unit name is
5643 always valid when services log to the journal via
5646 * There's a new man page kernel-command-line(7) detailing all
5647 command line options we understand.
5649 * The fstab generator may now be disabled at boot by passing
5650 fstab=0 on the kernel command line.
5652 * A new kernel command line option modules-load= is now understood
5653 to load a specific kernel module statically, early at boot.
5655 * Unit names specified on the systemctl command line are now
5656 automatically escaped as needed. Also, if file system or
5657 device paths are specified they are automatically turned
5658 into the appropriate mount or device unit names. Example:
5660 systemctl status /home
5661 systemctl status /dev/sda
5663 * The SysVConsole= configuration option has been removed from
5664 system.conf parsing.
5666 * The SysV search path is no longer exported on the D-Bus
5669 * The Names= option is been removed from unit file parsing.
5671 * There's a new man page bootup(7) detailing the boot process.
5673 * Every unit and every generator we ship with systemd now
5674 comes with full documentation. The self-explanatory boot is
5677 * A couple of services gained "systemd-" prefixes in their
5678 name if they wrap systemd code, rather than only external
5679 code. Among them fsck@.service which is now
5680 systemd-fsck@.service.
5682 * The HaveWatchdog property has been removed from the D-Bus
5685 * systemd.confirm_spawn= on the kernel command line should now
5688 * There's a new man page crypttab(5) which details all options
5689 we actually understand.
5691 * systemd-nspawn gained a new --capability= switch to pass
5692 additional capabilities to the container.
5694 * timedated will now read known NTP implementation unit names
5695 from /usr/lib/systemd/ntp-units.d/*.list,
5696 systemd-timedated-ntp.target has been removed.
5698 * journalctl gained a new switch "-b" that lists log data of
5699 the current boot only.
5701 * The notify socket is in the abstract namespace again, in
5702 order to support daemons which chroot() at start-up.
5704 * There is a new Storage= configuration option for journald
5705 which allows configuration of where log data should go. This
5706 also provides a way to disable journal logging entirely, so
5707 that data collected is only forwarded to the console, the
5708 kernel log buffer or another syslog implementation.
5710 * Many bugfixes and optimizations
5712 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
5713 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
5714 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
5715 Shawn Landden, Tom Gundersen
5719 * "systemctl help <unit>" now shows the man page if one is
5722 * Several new man pages have been added.
5724 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
5725 MaxLevelConsole= can now be specified in
5726 journald.conf. These options allow reducing the amount of
5727 data stored on disk or forwarded by the log level.
5729 * TimerSlackNSec= can now be specified in system.conf for
5730 PID1. This allows system-wide power savings.
5732 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
5733 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
5738 * logind is now capable of (optionally) handling power and
5739 sleep keys as well as the lid switch.
5741 * journalctl now understands the syntax "journalctl
5742 /usr/bin/avahi-daemon" to get all log output of a specific
5745 * CapabilityBoundingSet= in system.conf now also influences
5746 the capability bound set of usermode helpers of the kernel.
5748 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
5749 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
5750 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
5751 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
5755 * Note that we skipped 139 releases here in order to set the
5756 new version to something that is greater than both udev's
5757 and systemd's most recent version number.
5759 * udev: all udev sources are merged into the systemd source tree now.
5760 All future udev development will happen in the systemd tree. It
5761 is still fully supported to use the udev daemon and tools without
5762 systemd running, like in initramfs or other init systems. Building
5763 udev though, will require the *build* of the systemd tree, but
5764 udev can be properly *run* without systemd.
5766 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
5767 should be used to create dead device nodes as workarounds for broken
5770 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
5771 no longer supported. udev_monitor_new_from_netlink() needs to be
5772 used to subscribe to events.
5774 * udev: when udevd is started by systemd, processes which are left
5775 behind by forking them off of udev rules, are unconditionally cleaned
5776 up and killed now after the event handling has finished. Services or
5777 daemons must be started as systemd services. Services can be
5778 pulled-in by udev to get started, but they can no longer be directly
5779 forked by udev rules.
5781 * udev: the daemon binary is called systemd-udevd now and installed
5782 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
5783 to adapt to that, create symlink, or rename the binary after building
5786 * libudev no longer provides these symbols:
5787 udev_monitor_from_socket()
5788 udev_queue_get_failed_list_entry()
5789 udev_get_{dev,sys,run}_path()
5790 The versions number was bumped and symbol versioning introduced.
5792 * systemd-loginctl and systemd-journalctl have been renamed
5793 to loginctl and journalctl to match systemctl.
5795 * The config files: /etc/systemd/systemd-logind.conf and
5796 /etc/systemd/systemd-journald.conf have been renamed to
5797 logind.conf and journald.conf. Package updates should rename
5798 the files to the new names on upgrade.
5800 * For almost all files the license is now LGPL2.1+, changed
5801 from the previous GPL2.0+. Exceptions are some minor stuff
5802 of udev (which will be changed to LGPL2.1 eventually, too),
5803 and the MIT licensed sd-daemon.[ch] library that is suitable
5804 to be used as drop-in files.
5806 * systemd and logind now handle system sleep states, in
5807 particular suspending and hibernating.
5809 * logind now implements a sleep/shutdown/idle inhibiting logic
5810 suitable for a variety of uses. Soonishly Lennart will blog
5811 about this in more detail.
5813 * var-run.mount and var-lock.mount are no longer provided
5814 (which prevously bind mounted these directories to their new
5815 places). Distributions which have not converted these
5816 directories to symlinks should consider stealing these files
5817 from git history and add them downstream.
5819 * We introduced the Documentation= field for units and added
5820 this to all our shipped units. This is useful to make it
5821 easier to explore the boot and the purpose of the various
5824 * All smaller setup units (such as
5825 systemd-vconsole-setup.service) now detect properly if they
5826 are run in a container and are skipped when
5827 appropriate. This guarantees an entirely noise-free boot in
5828 Linux container environments such as systemd-nspawn.
5830 * A framework for implementing offline system updates is now
5831 integrated, for details see:
5832 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
5834 * A new service type Type=idle is available now which helps us
5835 avoiding ugly interleaving of getty output and boot status
5838 * There's now a system-wide CapabilityBoundingSet= option to
5839 globally reduce the set of capabilities for the
5840 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
5841 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
5842 even CAP_NET_ADMIN system-wide for secure systems.
5844 * There are now system-wide DefaultLimitXXX= options to
5845 globally change the defaults of the various resource limits
5846 for all units started by PID 1.
5848 * Harald Hoyer's systemd test suite has been integrated into
5849 systemd which allows easy testing of systemd builds in qemu
5850 and nspawn. (This is really awesome! Ask us for details!)
5852 * The fstab parser is now implemented as generator, not inside
5855 * systemctl will now warn you if .mount units generated from
5856 /etc/fstab are out of date due to changes in fstab that
5857 have not been read by systemd yet.
5859 * systemd is now suitable for usage in initrds. Dracut has
5860 already been updated to make use of this. With this in place
5861 initrds get a slight bit faster but primarily are much
5862 easier to introspect and debug since "systemctl status" in
5863 the host system can be used to introspect initrd services,
5864 and the journal from the initrd is kept around too.
5866 * systemd-delta has been added, a tool to explore differences
5867 between user/admin configuration and vendor defaults.
5869 * PrivateTmp= now affects both /tmp and /var/tmp.
5871 * Boot time status messages are now much prettier and feature
5872 proper english language. Booting up systemd has never been
5875 * Read-ahead pack files now include the inode number of all
5876 files to pre-cache. When the inode changes the pre-caching
5877 is not attempted. This should be nicer to deal with updated
5878 packages which might result in changes of read-ahead
5881 * We now temporaritly lower the kernel's read_ahead_kb variable
5882 when collecting read-ahead data to ensure the kernel's
5883 built-in read-ahead does not add noise to our measurements
5884 of necessary blocks to pre-cache.
5886 * There's now RequiresMountsFor= to add automatic dependencies
5887 for all mounts necessary for a specific file system path.
5889 * MountAuto= and SwapAuto= have been removed from
5890 system.conf. Mounting file systems at boot has to take place
5893 * nspawn now learned a new switch --uuid= to set the machine
5894 ID on the command line.
5896 * nspawn now learned the -b switch to automatically search
5899 * vt102 is now the default TERM for serial TTYs, upgraded from
5902 * systemd-logind now works on VT-less systems.
5904 * The build tree has been reorganized. The individual
5905 components now have directories of their own.
5907 * A new condition type ConditionPathIsReadWrite= is now available.
5909 * nspawn learned the new -C switch to create cgroups for the
5910 container in other hierarchies.
5912 * We now have support for hardware watchdogs, configurable in
5915 * The scheduled shutdown logic now has a public API.
5917 * We now mount /tmp as tmpfs by default, but this can be
5918 masked and /etc/fstab can override it.
5920 * Since udisks does not make use of /media anymore we are not
5921 mounting a tmpfs on it anymore.
5923 * journalctl gained a new --local switch to only interleave
5924 locally generated journal files.
5926 * We can now load the IMA policy at boot automatically.
5928 * The GTK tools have been split off into a systemd-ui.
5930 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
5931 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
5932 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
5933 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
5934 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
5935 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
5936 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
5937 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
5938 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
5943 * This is mostly a bugfix release
5945 * Support optional initialization of the machine ID from the
5946 KVM or container configured UUID.
5948 * Support immediate reboots with "systemctl reboot -ff"
5950 * Show /etc/os-release data in systemd-analyze output
5952 * Many bugfixes for the journal, including endianness fixes and
5953 ensuring that disk space enforcement works
5955 * sd-login.h is C++ comptaible again
5957 * Extend the /etc/os-release format on request of the Debian
5960 * We now refuse non-UTF8 strings used in various configuration
5961 and unit files. This is done to ensure we do not pass invalid
5962 data over D-Bus or expose it elsewhere.
5964 * Register Mimo USB Screens as suitable for automatic seat
5967 * Read SELinux client context from journal clients in a race
5970 * Reorder configuration file lookup order. /etc now always
5971 overrides /run in order to allow the administrator to always
5972 and unconditionally override vendor-supplied or
5973 automatically generated data.
5975 * The various user visible bits of the journal now have man
5976 pages. We still lack man pages for the journal API calls
5979 * We now ship all man pages in HTML format again in the
5982 Contributions from: Dave Reisner, Dirk Eibach, Frederic
5983 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
5984 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
5989 * This is mostly a bugfix release
5991 * systems lacking /etc/os-release are no longer supported.
5993 * Various functionality updates to libsystemd-login.so
5995 * Track class of PAM logins to distinguish greeters from
5998 Contributions from: Kay Sievers, Lennart Poettering, Michael
6003 * This is an important bugfix release for v41.
6005 * Building man pages is now optional which should be useful
6006 for those building systemd from git but unwilling to install
6009 * Watchdog support for supervising services is now usable. In
6010 a future release support for hardware watchdogs
6011 (i.e. /dev/watchdog) will be added building on this.
6013 * Service start rate limiting is now configurable and can be
6014 turned off per service. When a start rate limit is hit a
6015 reboot can automatically be triggered.
6017 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
6019 Contributions from: Benjamin Franzke, Bill Nottingham,
6020 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
6021 Schmidt, Michał Górny, Piotr Drąg
6025 * The systemd binary is installed /usr/lib/systemd/systemd now;
6026 An existing /sbin/init symlink needs to be adapted with the
6029 * The code that loads kernel modules has been ported to invoke
6030 libkmod directly, instead of modprobe. This means we do not
6031 support systems with module-init-tools anymore.
6033 * Watchdog support is now already useful, but still not
6036 * A new kernel command line option systemd.setenv= is
6037 understood to set system wide environment variables
6038 dynamically at boot.
6040 * We now limit the set of capabilities of systemd-journald.
6042 * We now set SIGPIPE to ignore by default, since it only is
6043 useful in shell pipelines, and has little use in general
6044 code. This can be disabled with IgnoreSIPIPE=no in unit
6047 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
6048 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
6053 * This is mostly a bugfix release
6055 * We now expose the reason why a service failed in the
6056 "Result" D-Bus property.
6058 * Rudimentary service watchdog support (will be completed over
6059 the next few releases.)
6061 * When systemd forks off in order execute some service we will
6062 now immediately changes its argv[0] to reflect which process
6063 it will execute. This is useful to minimize the time window
6064 with a generic argv[0], which makes bootcharts more useful
6066 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
6067 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
6068 Mike Kazantsev, Ray Strode
6072 * This is mostly a test release, but incorporates many
6075 * New systemd-cgtop tool to show control groups by their
6078 * Linking against libacl for ACLs is optional again. If
6079 disabled, support tracking device access for active logins
6080 goes becomes unavailable, and so does access to the user
6081 journals by the respective users.
6083 * If a group "adm" exists, journal files are automatically
6084 owned by them, thus allow members of this group full access
6085 to the system journal as well as all user journals.
6087 * The journal now stores the SELinux context of the logging
6088 client for all entries.
6090 * Add C++ inclusion guards to all public headers
6092 * New output mode "cat" in the journal to print only text
6093 messages, without any meta data like date or time.
6095 * Include tiny X server wrapper as a temporary stop-gap to
6096 teach XOrg udev display enumeration. This is used by display
6097 managers such as gdm, and will go away as soon as XOrg
6098 learned native udev hotplugging for display devices.
6100 * Add new systemd-cat tool for executing arbitrary programs
6101 with STDERR/STDOUT connected to the journal. Can also act as
6102 BSD logger replacement, and does so by default.
6104 * Optionally store all locally generated coredumps in the
6105 journal along with meta data.
6107 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
6108 writing short strings to files (for usage for /sys), and for
6109 creating symlinks, character and block device nodes.
6111 * New unit file option ControlGroupPersistent= to make cgroups
6112 persistent, following the mechanisms outlined in
6113 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
6115 * Support multiple local RTCs in a sane way
6117 * No longer monopolize IO when replaying readahead data on
6118 rotating disks, since we might starve non-file-system IO to
6119 death, since fanotify() will not see accesses done by blkid,
6122 * Do not show kernel threads in systemd-cgls anymore, unless
6123 requested with new -k switch.
6125 Contributions from: Dan Horák, Kay Sievers, Lennart
6126 Poettering, Michal Schmidt
6130 * This is mostly a test release, but incorporates many
6133 * The git repository moved to:
6134 git://anongit.freedesktop.org/systemd/systemd
6135 ssh://git.freedesktop.org/git/systemd/systemd
6137 * First release with the journal
6138 http://0pointer.de/blog/projects/the-journal.html
6140 * The journal replaces both systemd-kmsg-syslogd and
6141 systemd-stdout-bridge.
6143 * New sd_pid_get_unit() API call in libsystemd-logind
6145 * Many systemadm clean-ups
6147 * Introduce remote-fs-pre.target which is ordered before all
6148 remote mounts and may be used to start services before all
6151 * Added Mageia support
6153 * Add bash completion for systemd-loginctl
6155 * Actively monitor PID file creation for daemons which exit in
6156 the parent process before having finished writing the PID
6157 file in the daemon process. Daemons which do this need to be
6158 fixed (i.e. PID file creation must have finished before the
6159 parent exits), but we now react a bit more gracefully to them.
6161 * Add colourful boot output, mimicking the well-known output
6162 of existing distributions.
6164 * New option PassCredentials= for socket units, for
6165 compatibility with a recent kernel ABI breakage.
6167 * /etc/rc.local is now hooked in via a generator binary, and
6168 thus will no longer act as synchronization point during
6171 * systemctl list-unit-files now supports --root=.
6173 * systemd-tmpfiles now understands two new commands: z, Z for
6174 relabelling files according to the SELinux database. This is
6175 useful to apply SELinux labels to specific files in /sys,
6178 * Output of SysV services is now forwarded to both the console
6179 and the journal by default, not only just the console.
6181 * New man pages for all APIs from libsystemd-login.
6183 * The build tree got reorganized and a the build system is a
6184 lot more modular allowing embedded setups to specifically
6185 select the components of systemd they are interested in.
6187 * Support for Linux systems lacking the kernel VT subsystem is
6190 * configure's --with-rootdir= got renamed to
6191 --with-rootprefix= to follow the naming used by udev and
6194 * Unless specified otherwise we will now install to /usr instead
6195 of /usr/local by default.
6197 * Processes with '@' in argv[0][0] are now excluded from the
6198 final shut-down killing spree, following the logic explained
6200 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
6202 * All processes remaining in a service cgroup when we enter
6203 the START or START_PRE states are now killed with
6204 SIGKILL. That means it is no longer possible to spawn
6205 background processes from ExecStart= lines (which was never
6206 supported anyway, and bad style).
6208 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
6209 reloading of units together.
6211 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
6212 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
6213 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
6214 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
6215 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek