1 systemd System and Service Manager
5 * In service units the various ExecXYZ= settings have been extended
6 with an additional special character as first argument of the
7 assigned value: if the character '+' is used the specified command
8 line it will be run with full privileges, regardless of User=,
9 Group=, CapabilityBoundingSet= and similar options. The effect is
10 similar to the existing PermissionsStartOnly= option, but allows
11 configuration of this concept for each executed command line
14 * Services may now alter the service watchdog timeout at runtime by
15 sending a WATCHDOG_USEC= message via sd_notify().
17 * MemoryLimit= and related unit settings now optionally take percentage
18 specifications. The percentage is taken relative to the amount of
19 physical memory in the system (or in case of containers, the assigned
20 amount of memory). This allows scaling service resources neatly with
21 the amount of RAM available on the system. Similarly, systemd-logind's
22 RuntimeDirectorySize= option now also optionally takes percentage
25 * In similar fashion TasksMax= takes percentage values now, too. The
26 value is taken relative to the configured maximum number of processes
27 on the system. The per-service task maximum has been changed to 15%
28 using this functionality. (Effectively this is an increase of 512 →
29 4915 for service units, given the kernel's default pid_max setting.)
31 * Calendar time specifications in .timer units now understand a ".."
32 syntax for time ranges. Example: "4..7:10" may now be used for
33 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
36 * The InaccessableDirectories=, ReadOnlyDirectories= and
37 ReadWriteDirectories= unit file settings have been renamed to
38 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
39 applied to all kinds of file nodes, and not just directories, with
40 the exception of symlinks. Specifically these settings may now be
41 used on block and character device nodes, UNIX sockets and FIFOS as
42 well as regular files. The old names of these settings remain
43 available for compatibility.
45 * systemd will now log about all service processes it kills forcibly
46 (using SIGKILL) because they remained after the clean shutdown phase
47 of the service completed. This should help identifying services that
48 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
49 systemd-logind's configuration a similar log message is generated for
50 processes killed at the end of each session due to this setting.
52 * systemd will now set the $JOURNAL_STREAM environment variable for all
53 services whose stdout/stderr are connected to the Journal (which
54 effectively means by default: all services). The variable contains
55 the device and inode number of the file descriptor used for
56 stdout/stderr. This may be used by invoked programs to detect whether
57 their stdout/stderr is connected to the Journal, in which case they
58 can switch over to direct Journal communication, thus being able to
59 pass extended, structured metadata along with their log messages. As
60 one example, this is now used by glib's logging primitives.
62 * When using systemd's default tmp.mount unit for /tmp, the mount point
63 will now be established with the "nosuid" and "nodev" options. This
64 avoids privilege escalation attacks that put traps and exploits into
65 /tmp. However, this might cause problems if you e. g. put container
66 images or overlays into /tmp; if you need this, override tmp.mount's
67 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
70 * systemd now supports the "memory" cgroup controller also on
73 * The systemd-cgtop tool now optionally takes a control group path as
74 command line argument. If specified, the control group list shown is
75 limited to subgroups of that group.
77 * The SystemCallFilter= unit file setting gained support for
78 pre-defined, named system call filter sets. For example
79 SystemCallFilter=@clock is now an effective way to make all clock
80 changing-related system calls unavailable to a service. A number of
81 similar pre-defined groups are defined. Writing system call filters
82 for system services is simplified substantially with this new
83 concept. Accordingly, all of systemd's own, long-running services now
84 enable system call filtering based on this, by default.
86 * A new service setting MemoryDenyWriteExecute= has been added, taking
87 a boolean value. If turned on, a service may no longer create memory
88 mappings that are writable and executable at the same time. This
89 enhances security for services where this is enabled as it becomes
90 harder to dynamically write and then execute memory in exploited
91 service processes. This option has been enabled for all of systemd's
92 own long-running services.
94 * A new RestrictRealtime= service setting has been added, taking a
95 boolean argument. If set the service's processes may no longer
96 acquire realtime scheduling. This improves security as realtime
97 scheduling may otherwise be used to easily freeze the system.
99 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
100 value. This may be used for requesting that the system manager inside
101 of the container reports start-up completion to nspawn which then
102 propagates this notification further to the service manager
103 supervising nspawn itself. A related option NotifyReady= in .nspawn
104 files has been added too. This functionality allows ordering of the
105 start-up of multiple containers using the usual systemd ordering
108 * machinectl gained a new command "stop" that is an alias for
111 * systemd-resolved gained support for contacting DNS servers on
112 link-local IPv6 addresses.
114 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
115 its caches. A method call for requesting the same operation has been
116 added to the bus API too, and is made available via "systemd-resolve
119 * systemd-resolve gained a new --status switch. If passed a brief
120 summary of the used DNS configuration with per-interface information
123 * resolved.conf gained a new Cache= boolean option, defaulting to
124 on. If turned off local DNS caching is disabled. This comes with a
125 performance penalty in particular when DNSSEC is enabled. Note that
126 resolved disables its internal caching implicitly anyway, when the
127 configured DNS server is on a host-local IP address such as ::1 or
128 127.0.0.1, thus automatically avoiding double local caching.
130 * systemd-resolved now listens on the local IP address 127.0.0.53:53
131 for DNS requests. This improves compatibility with local programs
132 that do not use the libc NSS or systemd-resolved's bus APIs for name
133 resolution. This minimal DNS service is only available to local
134 programs and does not implement the full DNS protocol, but enough to
135 cover local DNS clients. A new, static resolv.conf file, listing just
136 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
137 now recommended to make /etc/resolv.conf a symlink to this file in
138 order to route all DNS lookups to systemd-resolved, regardless if
139 done via NSS, the bus API or raw DNS packets. Note that this local
140 DNS service is not as fully featured as the libc NSS or
141 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
142 used to deliver link-local address information (as this implies
143 sending a local interface index along), LLMNR/mDNS support via this
144 interface is severely restricted. It is thus strongly recommended for
145 all applications to use the libc NSS API or native systemd-resolved
148 * systemd-networkd's bridge support learned a new setting
149 VLANFiltering= for controlling VLAN filtering. Moreover a new section
150 in .network files has been added for configuring VLAN bridging in
151 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
153 * systemd-networkd's IPv6 Router Advertisement code now makes use of
154 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
155 now be acquired without relying on DHCPv6. Two new options
156 UseDomains= and UseDNS= have been added to configure this behaviour.
158 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
159 renamed IPv6AcceptRA=, without altering its behaviour. The old
160 setting name remains available for compatibility reasons.
162 * The systemd-networkd VTI/VTI6 tunneling support gained new options
163 Key=, InputKey= and OutputKey=.
165 * systemd-networkd gained support for VRF ("Virtual Routing Function")
166 interface configuration.
168 * "systemctl edit" may now be used to create new unit files by
169 specifying the --force switch.
171 * sd-event gained a new function sd_event_get_iteration() for
172 requesting the current iteration counter of the event loop. It starts
173 at zero and is increased by one with each event loop iteration.
175 * A new rpm macro %systemd_ordering is provided by the macros.systemd
176 file. It can be used in lieu of %systemd_requires in packages which
177 don't use any systemd functionality and are intended to be installed
178 in minimal containers without systemd present. This macro provides
179 ordering dependecies to ensure that if the package is installed in
180 the same rpm transaction as systemd, systemd will be installed before
181 the scriptlets for the package are executed, allowing unit presets
184 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
185 been added to simplify packaging of generators.
187 * The os-release file gained VERSION_CODENAME field for the
188 distribution nickname (e.g. VERSION_CODENAME=woody).
190 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
191 can be set to disable parsing of metadata and the creation
192 of persistent symlinks for that device.
194 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
195 to make them available to logged-in users has been reverted.
197 * Much of the common code of the various systemd components is now
198 built into an internal shared library libsystemd-shared-231.so
199 (incorporating the systemd version number in the name, to be updated
200 with future releases) that the components link to. This should
201 decrease systemd footprint both in memory during runtime and on
202 disk. Note that the shared library is not for public use, and is
203 neither API not ABI stable, but is likely to change with every new
204 released update. Packagers need to make sure that binaries
205 linking to libsystemd-shared.so are updated in step with the
208 * Configuration for "mkosi" is now part of the systemd
209 repository. mkosi is a tool to easily build legacy-free OS images,
210 and is available on github: https://github.com/systemd/mkosi. If
211 "mkosi" is invoked in the build tree a new raw OS image is generated
212 incorporating the systemd sources currently being worked on and a
213 clean, fresh distribution installation. The generated OS image may be
214 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physcial
215 UEFI PC. This functionality is particularly useful to easily test
216 local changes made to systemd in a pristine, defined environment. See
219 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
220 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
221 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
222 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
223 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
224 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
225 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
226 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
227 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
228 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
229 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
230 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
231 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
232 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
233 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
234 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
235 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
236 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
237 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
243 * DNSSEC is now turned on by default in systemd-resolved (in
244 "allow-downgrade" mode), but may be turned off during compile time by
245 passing "--with-default-dnssec=no" to "configure" (and of course,
246 during runtime with DNSSEC= in resolved.conf). We recommend
247 downstreams to leave this on at least during development cycles and
248 report any issues with the DNSSEC logic upstream. We are very
249 interested in collecting feedback about the DNSSEC validator and its
250 limitations in the wild. Note however, that DNSSEC support is
251 probably nothing downstreams should turn on in stable distros just
252 yet, as it might create incompatibilities with a few DNS servers and
253 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
254 automatically whenever we detect such incompatible setups, but there
255 might be systems we do not cover yet. Hence: please help us testing
256 the DNSSEC code, leave this on where you can, report back, but then
257 again don't consider turning this on in your stable, LTS or
258 production release just yet. (Note that you have to enable
259 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
260 and its DNSSEC mode for host name resolution from local
263 * systemd-resolve conveniently resolves DANE records with the --tlsa
264 option and OPENPGPKEY records with the --openpgp option. It also
265 supports dumping raw DNS record data via the new --raw= switch.
267 * systemd-logind will now by default terminate user processes that are
268 part of the user session scope unit (session-XX.scope) when the user
269 logs out. This behavior is controlled by the KillUserProcesses=
270 setting in logind.conf, and the previous default of "no" is now
271 changed to "yes". This means that user sessions will be properly
272 cleaned up after, but additional steps are necessary to allow
273 intentionally long-running processes to survive logout.
275 While the user is logged in at least once, user@.service is running,
276 and any service that should survive the end of any individual login
277 session can be started at a user service or scope using systemd-run.
278 systemd-run(1) man page has been extended with an example which shows
279 how to run screen in a scope unit underneath user@.service. The same
280 command works for tmux.
282 After the user logs out of all sessions, user@.service will be
283 terminated too, by default, unless the user has "lingering" enabled.
284 To effectively allow users to run long-term tasks even if they are
285 logged out, lingering must be enabled for them. See loginctl(1) for
286 details. The default polkit policy was modified to allow users to
287 set lingering for themselves without authentication.
289 Previous defaults can be restored at compile time by the
290 --without-kill-user-processes option to "configure".
292 * systemd-logind gained new configuration settings SessionsMax= and
293 InhibitorsMax=, both with a default of 8192. It will not register new
294 user sessions or inhibitors above this limit.
296 * systemd-logind will now reload configuration on SIGHUP.
298 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
299 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
300 enable. Also, support for the "io" cgroup controller in the unified
301 hierarchy has been added, so that the "memory", "pids" and "io" are
302 now the controllers that are supported on the unified hierarchy.
304 WARNING: it is not possible to use previous systemd versions with
305 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
306 is necessary to also update systemd in the initramfs if using the
307 unified hierarchy. An updated SELinux policy is also required.
309 * LLDP support has been extended, and both passive (receive-only) and
310 active (sender) modes are supported. Passive mode ("routers-only") is
311 enabled by default in systemd-networkd. Active LLDP mode is enabled
312 by default for containers on the internal network. The "networkctl
313 lldp" command may be used to list information gathered. "networkctl
314 status" will also show basic LLDP information on connected peers now.
316 * The IAID and DUID unique identifier sent in DHCP requests may now be
317 configured for the system and each .network file managed by
318 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
320 * systemd-networkd gained support for configuring proxy ARP support for
321 each interface, via the ProxyArp= setting in .network files. It also
322 gained support for configuring the multicast querier feature of
323 bridge devices, via the new MulticastQuerier= setting in .netdev
324 files. Similarly, snooping on the IGMP traffic can be controlled
325 via the new setting MulticastSnooping=.
327 A new setting PreferredLifetime= has been added for addresses
328 configured in .network file to configure the lifetime intended for an
331 The systemd-networkd DHCP server gained the option EmitRouter=, which
332 defaults to yes, to configure whether the DHCP Option 3 (Router)
335 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
336 systemd-socket-activate and installed into /usr/bin. It is now fully
339 * systemd-journald now uses separate threads to flush changes to disk
340 when closing journal files, thus reducing impact of slow disk I/O on
343 * The sd-journal API gained two new calls
344 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
345 can be used to open journal files using file descriptors instead of
346 file or directory paths. sd_journal_open_container() has been
347 deprecated, sd_journal_open_directory_fd() should be used instead
348 with the flag SD_JOURNAL_OS_ROOT.
350 * journalctl learned a new output mode "-o short-unix" that outputs log
351 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
352 UTC). It also gained support for a new --no-hostname setting to
353 suppress the hostname column in the family of "short" output modes.
355 * systemd-ask-password now optionally skips printing of the password to
356 stdout with --no-output which can be useful in scripts.
358 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
359 (devices tagged with ID_MAKER_TOOL) are now tagged with
360 "uaccess" and are available to logged in users.
362 * The DeviceAllow= unit setting now supports specifiers (with "%").
364 * "systemctl show" gained a new --value switch, which allows print a
365 only the contents of a specific unit property, without also printing
366 the property's name. Similar support was added to "show*" verbs
367 of loginctl and machinectl that output "key=value" lists.
369 * A new unit type "generated" was added for files dynamically generated
370 by generator tools. Similarly, a new unit type "transient" is used
371 for unit files created using the runtime API. "systemctl enable" will
372 refuse to operate on such files.
374 * A new command "systemctl revert" has been added that may be used to
375 revert to the vendor version of a unit file, in case local changes
376 have been made by adding drop-ins or overriding the unit file.
378 * "machinectl clean" gained a new verb to automatically remove all or
379 just hidden container images.
381 * systemd-tmpfiles gained support for a new line type "e" for emptying
382 directories, if they exist, without creating them if they don't.
384 * systemd-nspawn gained support for automatically patching the UID/GIDs
385 of the owners and the ACLs of all files and directories in a
386 container tree to match the UID/GID user namespacing range selected
387 for the container invocation. This mode is enabled via the new
388 --private-user-chown switch. It also gained support for automatically
389 choosing a free, previously unused UID/GID range when starting a
390 container, via the new --private-users=pick setting (which implies
391 --private-user-chown). Together, these options for the first time
392 make user namespacing for nspawn containers fully automatic and thus
393 deployable. The systemd-nspawn@.service template unit file has been
394 changed to use this functionality by default.
396 * systemd-nspawn gained a new --network-zone= switch, that allows
397 creating ad-hoc virtual Ethernet links between multiple containers,
398 that only exist as long as at least one container referencing them is
399 running. This allows easy connecting of multiple containers with a
400 common link that implements an Ethernet broadcast domain. Each of
401 these network "zones" may be named relatively freely by the user, and
402 may be referenced by any number of containers, but each container may
403 only reference one of these "zones". On the lower level, this is
404 implemented by an automatically managed bridge network interface for
405 each zone, that is created when the first container referencing its
406 zone is created and removed when the last one referencing its zone
409 * The default start timeout may now be configured on the kernel command
410 line via systemd.default_timeout_start_sec=. It was already
411 configurable via the DefaultTimeoutStartSec= option in
412 /etc/systemd/system.conf.
414 * Socket units gained a new TriggerLimitIntervalSec= and
415 TriggerLimitBurst= setting to configure a limit on the activation
416 rate of the socket unit.
418 * The LimitNICE= setting now optionally takes normal UNIX nice values
419 in addition to the raw integer limit value. If the specified
420 parameter is prefixed with "+" or "-" and is in the range -20..19 the
421 value is understood as UNIX nice value. If not prefixed like this it
422 is understood as raw RLIMIT_NICE limit.
424 * Note that the effect of the PrivateDevices= unit file setting changed
425 slightly with this release: the per-device /dev file system will be
426 mounted read-only from this version on, and will have "noexec"
427 set. This (minor) change of behavior might cause some (exceptional)
428 legacy software to break, when PrivateDevices=yes is set for its
429 service. Please leave PrivateDevices= off if you run into problems
432 * systemd-bootchart has been split out to a separate repository:
433 https://github.com/systemd/systemd-bootchart
435 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
436 merged into the kernel in its current form.
438 * The compatibility libraries libsystemd-daemon.so,
439 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
440 which have been deprecated since systemd-209 have been removed along
441 with the corresponding pkg-config files. All symbols provided by
442 those libraries are provided by libsystemd.so.
444 * The Capabilities= unit file setting has been removed (it is ignored
445 for backwards compatibility). AmbientCapabilities= and
446 CapabilityBoundingSet= should be used instead.
448 * A new special target has been added, initrd-root-device.target,
449 which creates a synchronization point for dependencies of the root
450 device in early userspace. Initramfs builders must ensure that this
451 target is now included in early userspace.
453 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
454 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
455 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
456 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
457 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
458 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
459 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
460 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
461 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
462 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
463 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
464 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
465 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
466 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
467 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
468 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
469 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
470 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
471 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
472 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
473 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
474 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
475 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
476 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
477 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
480 — Fairfax, 2016-05-21
484 * The systemd-resolved DNS resolver service has gained a substantial
485 set of new features, most prominently it may now act as a DNSSEC
486 validating stub resolver. DNSSEC mode is currently turned off by
487 default, but is expected to be turned on by default in one of the
488 next releases. For now, we invite everybody to test the DNSSEC logic
489 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
490 service also gained a full set of D-Bus interfaces, including calls
491 to configure DNS and DNSSEC settings per link (for use by external
492 network management software). systemd-resolved and systemd-networkd
493 now distinguish between "search" and "routing" domains. The former
494 are used to qualify single-label names, the latter are used purely
495 for routing lookups within certain domains to specific links.
496 resolved now also synthesizes RRs for all entries from /etc/hosts.
498 * The systemd-resolve tool (which is a client utility for
499 systemd-resolved) has been improved considerably and is now fully
500 supported and documented. Hence it has moved from /usr/lib/systemd to
503 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
506 * The coredump collection logic has been reworked: when a coredump is
507 collected it is now written to disk, compressed and processed
508 (including stacktrace extraction) from a new instantiated service
509 systemd-coredump@.service, instead of directly from the
510 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
511 processing large coredumps can take up a substantial amount of
512 resources and time, and this previously happened entirely outside of
513 systemd's service supervision. With the new logic the core_pattern
514 hook only does minimal metadata collection before passing off control
515 to the new instantiated service, which is configured with a time
516 limit, a nice level and other settings to minimize negative impact on
517 the rest of the system. Also note that the new logic will honour the
518 RLIMIT_CORE setting of the crashed process, which now allows users
519 and processes to turn off coredumping for their processes by setting
522 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
523 and all forked processes by default. Previously, PID 1 would leave
524 the setting at "0" for all processes, as set by the kernel. Note that
525 the resource limit traditionally has no effect on the generated
526 coredumps on the system if the /proc/sys/kernel/core_pattern hook
527 logic is used. Since the limit is now honoured (see above) its
528 default has been changed so that the coredumping logic is enabled by
529 default for all processes, while allowing specific opt-out.
531 * When the stacktrace is extracted from processes of system users, this
532 is now done as "systemd-coredump" user, in order to sandbox this
533 potentially security sensitive parsing operation. (Note that when
534 processing coredumps of normal users this is done under the user ID
535 of process that crashed, as before.) Packagers should take notice
536 that it is now necessary to create the "systemd-coredump" system user
537 and group at package installation time.
539 * The systemd-activate socket activation testing tool gained support
540 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
541 and --seqpacket switches. It also has been extended to support both
542 new-style and inetd-style file descriptor passing. Use the new
543 --inetd switch to request inetd-style file descriptor passing.
545 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
546 variable, which takes a boolean value. If set to false, ANSI color
547 output is disabled in the tools even when run on a terminal that
550 * The VXLAN support in networkd now supports two new settings
551 DestinationPort= and PortRange=.
553 * A new systemd.machine_id= kernel command line switch has been added,
554 that may be used to set the machine ID in /etc/machine-id if it is
555 not initialized yet. This command line option has no effect if the
556 file is already initialized.
558 * systemd-nspawn gained a new --as-pid2 switch that invokes any
559 specified command line as PID 2 rather than PID 1 in the
560 container. In this mode PID 1 is a minimal stub init process that
561 implements the special POSIX and Linux semantics of PID 1 regarding
562 signal and child process management. Note that this stub init process
563 is implemented in nspawn itself and requires no support from the
564 container image. This new logic is useful to support running
565 arbitrary commands in the container, as normal processes are
566 generally not prepared to run as PID 1.
568 * systemd-nspawn gained a new --chdir= switch for setting the current
569 working directory for the process started in the container.
571 * "journalctl /dev/sda" will now output all kernel log messages for
572 specified device from the current boot, in addition to all devices
573 that are parents of it. This should make log output about devices
574 pretty useful, as long as kernel drivers attach enough metadata to
575 the log messages. (The usual SATA drivers do.)
577 * The sd-journal API gained two new calls
578 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
579 that report whether log data from /run or /var has been found.
581 * journalctl gained a new switch "--fields" that prints all journal
582 record field names currently in use in the journal. This is backed
583 by two new sd-journal API calls sd_journal_enumerate_fields() and
584 sd_journal_restart_fields().
586 * Most configurable timeouts in systemd now expect an argument of
587 "infinity" to turn them off, instead of "0" as before. The semantics
588 from now on is that a timeout of "0" means "now", and "infinity"
589 means "never". To maintain backwards compatibility, "0" continues to
590 turn off previously existing timeout settings.
592 * "systemctl reload-or-try-restart" has been renamed to "systemctl
593 try-reload-or-restart" to clarify what it actually does: the "try"
594 logic applies to both reloading and restarting, not just restarting.
595 The old name continues to be accepted for compatibility.
597 * On boot-up, when PID 1 detects that the system clock is behind the
598 release date of the systemd version in use, the clock is now set
599 to the latter. Previously, this was already done in timesyncd, in order
600 to avoid running with clocks set to the various clock epochs such as
601 1902, 1938 or 1970. With this change the logic is now done in PID 1
602 in addition to timesyncd during early boot-up, so that it is enforced
603 before the first process is spawned by systemd. Note that the logic
604 in timesyncd remains, as it is more comprehensive and ensures
605 clock monotonicity by maintaining a persistent timestamp file in
606 /var. Since /var is generally not available in earliest boot or the
607 initrd, this part of the logic remains in timesyncd, and is not done
610 * Support for tweaking details in net_cls.class_id through the
611 NetClass= configuration directive has been removed, as the kernel
612 people have decided to deprecate that controller in cgroup v2.
613 Userspace tools such as nftables are moving over to setting rules
614 that are specific to the full cgroup path of a task, which obsoletes
615 these controllers anyway. The NetClass= directive is kept around for
616 legacy compatibility reasons. For a more in-depth description of the
617 kernel change, please refer to the respective upstream commit:
619 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
621 * A new service setting RuntimeMaxSec= has been added that may be used
622 to specify a maximum runtime for a service. If the timeout is hit, the
623 service is terminated and put into a failure state.
625 * A new service setting AmbientCapabilities= has been added. It allows
626 configuration of additional Linux process capabilities that are
627 passed to the activated processes. This is only available on very
630 * The process resource limit settings in service units may now be used
631 to configure hard and soft limits individually.
633 * The various libsystemd APIs such as sd-bus or sd-event now publicly
634 expose support for gcc's __attribute__((cleanup())) C extension.
635 Specifically, for many object destructor functions alternative
636 versions have been added that have names suffixed with "p" and take a
637 pointer to a pointer to the object to destroy, instead of just a
638 pointer to the object itself. This is useful because these destructor
639 functions may be used directly as parameters to the cleanup
640 construct. Internally, systemd has been a heavy user of this GCC
641 extension for a long time, and with this change similar support is
642 now available to consumers of the library outside of systemd. Note
643 that by using this extension in your sources compatibility with old
644 and strictly ANSI compatible C compilers is lost. However, all gcc or
645 LLVM versions of recent years support this extension.
647 * Timer units gained support for a new setting RandomizedDelaySec= that
648 allows configuring some additional randomized delay to the configured
649 time. This is useful to spread out timer events to avoid load peaks in
650 clusters or larger setups.
652 * Calendar time specifications now support sub-second accuracy.
654 * Socket units now support listening on SCTP and UDP-lite protocol
657 * The sd-event API now comes with a full set of man pages.
659 * Older versions of systemd contained experimental support for
660 compressing journal files and coredumps with the LZ4 compressor that
661 was not compatible with the lz4 binary (due to API limitations of the
662 lz4 library). This support has been removed; only support for files
663 compatible with the lz4 binary remains. This LZ4 logic is now
664 officially supported and no longer considered experimental.
666 * The dkr image import logic has been removed again from importd. dkr's
667 micro-services focus doesn't fit into the machine image focus of
668 importd, and quickly got out of date with the upstream dkr API.
670 * Creation of the /run/lock/lockdev/ directory was dropped from
671 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
672 been available for many years. If you still need this, you need to
673 create your own tmpfiles.d config file with:
675 d /run/lock/lockdev 0775 root lock -
677 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
678 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
679 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
680 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
681 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
682 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
683 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
684 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
685 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
686 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
687 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
688 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
689 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
690 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
691 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
692 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
693 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
694 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
695 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
701 * A number of properties previously only settable in unit
702 files are now also available as properties to set when
703 creating transient units programmatically via the bus, as it
704 is exposed with systemd-run's --property=
705 setting. Specifically, these are: SyslogIdentifier=,
706 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
707 EnvironmentFile=, ReadWriteDirectories=,
708 ReadOnlyDirectories=, InaccessibleDirectories=,
709 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
711 * When creating transient services via the bus API it is now
712 possible to pass in a set of file descriptors to use as
713 STDIN/STDOUT/STDERR for the invoked process.
715 * Slice units may now be created transiently via the bus APIs,
716 similar to the way service and scope units may already be
719 * Wherever systemd expects a calendar timestamp specification
720 (like in journalctl's --since= and --until= switches) UTC
721 timestamps are now supported. Timestamps suffixed with "UTC"
722 are now considered to be in Universal Time Coordinated
723 instead of the local timezone. Also, timestamps may now
724 optionally be specified with sub-second accuracy. Both of
725 these additions also apply to recurring calendar event
726 specification, such as OnCalendar= in timer units.
728 * journalctl gained a new "--sync" switch that asks the
729 journal daemon to write all so far unwritten log messages to
730 disk and sync the files, before returning.
732 * systemd-tmpfiles learned two new line types "q" and "Q" that
733 operate like "v", but also set up a basic btrfs quota
734 hierarchy when used on a btrfs file system with quota
737 * tmpfiles' "v", "q" and "Q" will now create a plain directory
738 instead of a subvolume (even on a btrfs file system) if the
739 root directory is a plain directory, and not a
740 subvolume. This should simplify things with certain chroot()
741 environments which are not aware of the concept of btrfs
744 * systemd-detect-virt gained a new --chroot switch to detect
745 whether execution takes place in a chroot() environment.
747 * CPUAffinity= now takes CPU index ranges in addition to
750 * The various memory-related resource limit settings (such as
751 LimitAS=) now understand the usual K, M, G, ... suffixes to
752 the base of 1024 (IEC). Similar, the time-related resource
753 limit settings understand the usual min, h, day, ...
756 * There's a new system.conf setting DefaultTasksMax= to
757 control the default TasksMax= setting for services and
758 scopes running on the system. (TasksMax= is the primary
759 setting that exposes the "pids" cgroup controller on systemd
760 and was introduced in the previous systemd release.) The
761 setting now defaults to 512, which means services that are
762 not explicitly configured otherwise will only be able to
763 create 512 processes or threads at maximum, from this
764 version on. Note that this means that thread- or
765 process-heavy services might need to be reconfigured to set
766 TasksMax= to a higher value. It is sufficient to set
767 TasksMax= in these specific unit files to a higher value, or
768 even "infinity". Similar, there's now a logind.conf setting
769 UserTasksMax= that defaults to 4096 and limits the total
770 number of processes or tasks each user may own
771 concurrently. nspawn containers also have the TasksMax=
772 value set by default now, to 8192. Note that all of this
773 only has an effect if the "pids" cgroup controller is
774 enabled in the kernel. The general benefit of these changes
775 should be a more robust and safer system, that provides a
776 certain amount of per-service fork() bomb protection.
778 * systemd-nspawn gained the new --network-veth-extra= switch
779 to define additional and arbitrarily-named virtual Ethernet
780 links between the host and the container.
782 * A new service execution setting PassEnvironment= has been
783 added that allows importing select environment variables
784 from PID1's environment block into the environment block of
787 * Timer units gained support for a new RemainAfterElapse=
788 setting which takes a boolean argument. It defaults to on,
789 exposing behaviour unchanged to previous releases. If set to
790 off, timer units are unloaded after they elapsed if they
791 cannot elapse again. This is particularly useful for
792 transient timer units, which shall not stay around longer
793 than until they first elapse.
795 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
796 default now (the kernel default is 16). This is beneficial
797 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
798 allows substantially larger numbers of queued
799 datagrams. This should increase the capability of systemd to
800 parallelize boot-up, as logging and sd_notify() are unlikely
801 to stall execution anymore. If you need to change the value
802 from the new defaults, use the usual sysctl.d/ snippets.
804 * The compression framing format used by the journal or
805 coredump processing has changed to be in line with what the
806 official LZ4 tools generate. LZ4 compression support in
807 systemd was considered unsupported previously, as the format
808 was not compatible with the normal tools. With this release
809 this has changed now, and it is hence safe for downstream
810 distributions to turn it on. While not compressing as well
811 as the XZ, LZ4 is substantially faster, which makes
812 it a good default choice for the compression logic in the
813 journal and in coredump handling.
815 * Any reference to /etc/mtab has been dropped from
816 systemd. The file has been obsolete since a while, but
817 systemd refused to work on systems where it was incorrectly
818 set up (it should be a symlink or non-existent). Please make
819 sure to update to util-linux 2.27.1 or newer in conjunction
820 with this systemd release, which also drops any reference to
821 /etc/mtab. If you maintain a distribution make sure that no
822 software you package still references it, as this is a
823 likely source of bugs. There's also a glibc bug pending,
824 asking for removal of any reference to this obsolete file:
826 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
828 Note that only util-linux versions built with
829 --enable-libmount-force-mountinfo are supported.
831 * Support for the ".snapshot" unit type has been removed. This
832 feature turned out to be little useful and little used, and
833 has now been removed from the core and from systemctl.
835 * The dependency types RequiresOverridable= and
836 RequisiteOverridable= have been removed from systemd. They
837 have been used only very sparingly to our knowledge and
838 other options that provide a similar effect (such as
839 systemctl --mode=ignore-dependencies) are much more useful
840 and commonly used. Moreover, they were only half-way
841 implemented as the option to control behaviour regarding
842 these dependencies was never added to systemctl. By removing
843 these dependency types the execution engine becomes a bit
844 simpler. Unit files that use these dependencies should be
845 changed to use the non-Overridable dependency types
846 instead. In fact, when parsing unit files with these
847 options, that's what systemd will automatically convert them
848 too, but it will also warn, asking users to fix the unit
849 files accordingly. Removal of these dependency types should
850 only affect a negligible number of unit files in the wild.
852 * Behaviour of networkd's IPForward= option changed
853 (again). It will no longer maintain a per-interface setting,
854 but propagate one way from interfaces where this is enabled
855 to the global kernel setting. The global setting will be
856 enabled when requested by a network that is set up, but
857 never be disabled again. This change was made to make sure
858 IPv4 and IPv6 behaviour regarding packet forwarding is
859 similar (as the Linux IPv6 stack does not support
860 per-interface control of this setting) and to minimize
863 * In unit files the behaviour of %u, %U, %h, %s has
864 changed. These specifiers will now unconditionally resolve
865 to the various user database fields of the user that the
866 systemd instance is running as, instead of the user
867 configured in the specific unit via User=. Note that this
868 effectively doesn't change much, as resolving of these
869 specifiers was already turned off in the --system instance
870 of systemd, as we cannot do NSS lookups from PID 1. In the
871 --user instance of systemd these specifiers where correctly
872 resolved, but hardly made any sense, since the user instance
873 lacks privileges to do user switches anyway, and User= is
874 hence useless. Morever, even in the --user instance of
875 systemd behaviour was awkward as it would only take settings
876 from User= assignment placed before the specifier into
877 account. In order to unify and simplify the logic around
878 this the specifiers will now always resolve to the
879 credentials of the user invoking the manager (which in case
880 of PID 1 is the root user).
882 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
883 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
884 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
885 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
886 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
887 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
888 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
889 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
890 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
891 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
892 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
893 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
894 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
895 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
902 * systemd now depends on util-linux v2.27. More specifically,
903 the newly added mount monitor feature in libmount now
904 replaces systemd's former own implementation.
906 * libmount mandates /etc/mtab not to be regular file, and
907 systemd now enforces this condition at early boot.
908 /etc/mtab has been deprecated and warned about for a very
909 long time, so systems running systemd should already have
910 stopped having this file around as anything else than a
911 symlink to /proc/self/mounts.
913 * Support for the "pids" cgroup controller has been added. It
914 allows accounting the number of tasks in a cgroup and
915 enforcing limits on it. This adds two new setting
916 TasksAccounting= and TasksMax= to each unit, as well as a
917 global option DefaultTasksAccounting=.
919 * Support for the "net_cls" cgroup controller has been added.
920 It allows assigning a net class ID to each task in the
921 cgroup, which can then be used in firewall rules and traffic
922 shaping configurations. Note that the kernel netfilter net
923 class code does not currently work reliably for ingress
924 packets on unestablished sockets.
926 This adds a new config directive called NetClass= to CGroup
927 enabled units. Allowed values are positive numbers for fixed
928 assignments and "auto" for picking a free value
931 * 'systemctl is-system-running' now returns 'offline' if the
932 system is not booted with systemd. This command can now be
933 used as a substitute for 'systemd-notify --booted'.
935 * Watchdog timeouts have been increased to 3 minutes for all
936 in-tree service files. Apparently, disk IO issues are more
937 frequent than we hoped, and user reported >1 minute waiting
940 * 'machine-id-commit' functionality has been merged into
941 'machine-id-setup --commit'. The separate binary has been
944 * The WorkingDirectory= directive in unit files may now be set
945 to the special value '~'. In this case, the working
946 directory is set to the home directory of the user
949 * "machinectl shell" will now open the shell in the home
950 directory of the selected user by default.
952 * The CrashChVT= configuration file setting is renamed to
953 CrashChangeVT=, following our usual logic of not
954 abbreviating unnecessarily. The old directive is still
955 supported for compat reasons. Also, this directive now takes
956 an integer value between 1 and 63, or a boolean value. The
957 formerly supported '-1' value for disabling stays around for
960 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
961 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
962 RootDirectory= properties can now be set for transient
965 * The systemd-analyze tool gained a new "set-log-target" verb
966 to change the logging target the system manager logs to
967 dynamically during runtime. This is similar to how
968 "systemd-analyze set-log-level" already changes the log
971 * In nspawn /sys is now mounted as tmpfs, with only a selected
972 set of subdirectories mounted in from the real sysfs. This
973 enhances security slightly, and is useful for ensuring user
974 namespaces work correctly.
976 * Support for USB FunctionFS activation has been added. This
977 allows implementation of USB gadget services that are
978 activated as soon as they are requested, so that they don't
979 have to run continuously, similar to classic socket
982 * The "systemctl exit" command now optionally takes an
983 additional parameter that sets the exit code to return from
984 the systemd manager when exiting. This is only relevant when
985 running the systemd user instance, or when running the
986 system instance in a container.
988 * sd-bus gained the new API calls sd_bus_path_encode_many()
989 and sd_bus_path_decode_many() that allow easy encoding and
990 decoding of multiple identifier strings inside a D-Bus
991 object path. Another new call sd_bus_default_flush_close()
992 has been added to flush and close per-thread default
995 * systemd-cgtop gained support for a -M/--machine= switch to
996 show the control groups within a certain container only.
998 * "systemctl kill" gained support for an optional --fail
999 switch. If specified the requested operation will fail of no
1000 processes have been killed, because the unit had no
1001 processes attached, or similar.
1003 * A new systemd.crash_reboot=1 kernel command line option has
1004 been added that triggers a reboot after crashing. This can
1005 also be set through CrashReboot= in systemd.conf.
1007 * The RuntimeDirectory= setting now understands unit
1008 specifiers like %i or %f.
1010 * A new (still internal) libary API sd-ipv4acd has been added,
1011 that implements address conflict detection for IPv4. It's
1012 based on code from sd-ipv4ll, and will be useful for
1013 detecting DHCP address conflicts.
1015 * File descriptors passed during socket activation may now be
1016 named. A new API sd_listen_fds_with_names() is added to
1017 access the names. The default names may be overridden,
1018 either in the .socket file using the FileDescriptorName=
1019 parameter, or by passing FDNAME= when storing the file
1020 descriptors using sd_notify().
1022 * systemd-networkd gained support for:
1024 - Setting the IPv6 Router Advertisement settings via
1025 IPv6AcceptRouterAdvertisements= in .network files.
1027 - Configuring the HelloTimeSec=, MaxAgeSec= and
1028 ForwardDelaySec= bridge parameters in .netdev files.
1030 - Configuring PreferredSource= for static routes in
1033 * The "ask-password" framework used to query for LUKS harddisk
1034 passwords or SSL passwords during boot gained support for
1035 caching passwords in the kernel keyring, if it is
1036 available. This makes sure that the user only has to type in
1037 a passphrase once if there are multiple objects to unlock
1038 with the same one. Previously, such password caching was
1039 available only when Plymouth was used; this moves the
1040 caching logic into the systemd codebase itself. The
1041 "systemd-ask-password" utility gained a new --keyname=
1042 switch to control which kernel keyring key to use for
1043 caching a password in. This functionality is also useful for
1044 enabling display managers such as gdm to automatically
1045 unlock the user's GNOME keyring if its passphrase, the
1046 user's password and the harddisk password are the same, if
1047 gdm-autologin is used.
1049 * When downloading tar or raw images using "machinectl
1050 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
1051 file is now also downloaded, if it is available and stored
1052 next to the image file.
1054 * Units of type ".socket" gained a new boolean setting
1055 Writable= which is only useful in conjunction with
1056 ListenSpecial=. If true, enables opening the specified
1057 special file in O_RDWR mode rather than O_RDONLY mode.
1059 * systemd-rfkill has been reworked to become a singleton
1060 service that is activated through /dev/rfkill on each rfkill
1061 state change and saves the settings to disk. This way,
1062 systemd-rfkill is now compatible with devices that exist
1063 only intermittendly, and even restores state if the previous
1064 system shutdown was abrupt rather than clean.
1066 * The journal daemon gained support for vacuuming old journal
1067 files controlled by the number of files that shall remain,
1068 in addition to the already existing control by size and by
1069 date. This is useful as journal interleaving performance
1070 degrades with too many separate journal files, and allows
1071 putting an effective limit on them. The new setting defaults
1072 to 100, but this may be changed by setting SystemMaxFiles=
1073 and RuntimeMaxFiles= in journald.conf. Also, the
1074 "journalctl" tool gained the new --vacuum-files= switch to
1075 manually vacuum journal files to leave only the specified
1076 number of files in place.
1078 * udev will now create /dev/disk/by-path links for ATA devices
1079 on kernels where that is supported.
1081 * Galician, Serbian, Turkish and Korean translations were added.
1083 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
1084 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
1085 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
1086 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
1087 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
1088 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
1089 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
1090 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
1091 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
1092 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
1093 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
1094 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
1095 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
1096 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
1097 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
1098 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
1099 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
1100 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
1102 — Berlin, 2015-10-07
1106 * The DHCP implementation of systemd-networkd gained a set of
1109 - The DHCP server now supports emitting DNS and NTP
1110 information. It may be enabled and configured via
1111 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
1112 and NTP information is enabled, but no servers are
1113 configured, the corresponding uplink information (if there
1114 is any) is propagated.
1116 - Server and client now support transmission and reception
1117 of timezone information. It can be configured via the
1118 newly introduced network options UseTimezone=,
1119 EmitTimezone=, and Timezone=. Transmission of timezone
1120 information is enabled between host and containers by
1121 default now: the container will change its local timezone
1122 to what the host has set.
1124 - Lease timeouts can now be configured via
1125 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
1127 - The DHCP server improved on the stability of
1128 leases. Clients are more likely to get the same lease
1129 information back, even if the server loses state.
1131 - The DHCP server supports two new configuration options to
1132 control the lease address pool metrics, PoolOffset= and
1135 * The encapsulation limit of tunnels in systemd-networkd may
1136 now be configured via 'EncapsulationLimit='. It allows
1137 modifying the maximum additional levels of encapsulation
1138 that are permitted to be prepended to a packet.
1140 * systemd now supports the concept of user buses replacing
1141 session buses, if used with dbus-1.10 (and enabled via dbus
1142 --enable-user-session). It previously only supported this on
1143 kdbus-enabled systems, and this release expands this to
1144 'dbus-daemon' systems.
1146 * systemd-networkd now supports predictable interface names
1149 * systemd now optionally supports the new Linux kernel
1150 "unified" control group hierarchy. If enabled via the kernel
1151 command-line option 'systemd.unified_cgroup_hierarchy=1',
1152 systemd will try to mount the unified cgroup hierarchy
1153 directly on /sys/fs/cgroup. If not enabled, or not
1154 available, systemd will fall back to the legacy cgroup
1155 hierarchy setup, as before. Host system and containers can
1156 mix and match legacy and unified hierarchies as they
1157 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
1158 environment variable to individually select the hierarchy to
1159 use for executed containers. By default, nspawn will use the
1160 unified hierarchy for the containers if the host uses the
1161 unified hierarchy, and the legacy hierarchy otherwise.
1162 Please note that at this point the unified hierarchy is an
1163 experimental kernel feature and is likely to change in one
1164 of the next kernel releases. Therefore, it should not be
1165 enabled by default in downstream distributions yet. The
1166 minimum required kernel version for the unified hierarchy to
1167 work is 4.2. Note that when the unified hierarchy is used
1168 for the first time delegated access to controllers is
1169 safe. Because of this systemd-nspawn containers will get
1170 access to controllers now, as will systemd user
1171 sessions. This means containers and user sessions may now
1172 manage their own resources, partitioning up what the system
1175 * A new special scope unit "init.scope" has been introduced
1176 that encapsulates PID 1 of the system. It may be used to
1177 determine resource usage and enforce resource limits on PID
1178 1 itself. PID 1 hence moved out of the root of the control
1181 * The cgtop tool gained support for filtering out kernel
1182 threads when counting tasks in a control group. Also, the
1183 count of processes is now recursively summed up by
1184 default. Two options -k and --recursive= have been added to
1185 revert to old behaviour. The tool has also been updated to
1186 work correctly in containers now.
1188 * systemd-nspawn's --bind= and --bind-ro= options have been
1189 extended to allow creation of non-recursive bind mounts.
1191 * libsystemd gained two new calls sd_pid_get_cgroup() and
1192 sd_peer_get_cgroup() which return the control group path of
1193 a process or peer of a connected AF_UNIX socket. This
1194 function call is particularly useful when implementing
1195 delegated subtrees support in the control group hierarchy.
1197 * The "sd-event" event loop API of libsystemd now supports
1198 correct dequeuing of real-time signals, without losing
1201 * When systemd requests a PolicyKit decision when managing
1202 units it will now add additional fields to the request,
1203 including unit name and desired operation. This enables more
1204 powerful PolicyKit policies, that make decisions depending
1205 on these parameters.
1207 * nspawn learnt support for .nspawn settings files, that may
1208 accompany the image files or directories of containers, and
1209 may contain additional settings for the container. This is
1210 an alternative to configuring container parameters via the
1211 nspawn command line.
1213 Contributions from: Cristian Rodríguez, Daniel Mack, David
1214 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
1215 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
1216 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
1217 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
1218 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
1219 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
1220 Andersen, Tom Gundersen, Torstein Husebø
1222 — Berlin, 2015-09-08
1226 * machinectl gained a new verb 'shell' which opens a fresh
1227 shell on the target container or the host. It is similar to
1228 the existing 'login' command of machinectl, but spawns the
1229 shell directly without prompting for username or
1230 password. The pseudo machine '.host' now refers to the local
1231 host and is used by default. Hence, 'machinectl shell' can
1232 be used as replacement for 'su -' which spawns a session as
1233 a fresh systemd unit in a way that is fully isolated from
1234 the originating session.
1236 * systemd-networkd learned to cope with private-zone DHCP
1237 options and allows other programs to query the values.
1239 * SELinux access control when enabling/disabling units is no
1240 longer enforced with this release. The previous
1241 implementation was incorrect, and a new corrected
1242 implementation is not yet available. As unit file operations
1243 are still protected via PolicyKit and D-Bus policy this is
1244 not a security problem. Yet, distributions which care about
1245 optimal SELinux support should probably not stabilize on
1248 * sd-bus gained support for matches of type "arg0has=", that
1249 test for membership of strings in string arrays sent in bus
1252 * systemd-resolved now dumps the contents of its DNS and LLMNR
1253 caches to the logs on reception of the SIGUSR1 signal. This
1254 is useful to debug DNS behaviour.
1256 * The coredumpctl tool gained a new --directory= option to
1257 operate on journal files in a specific directory.
1259 * "systemctl reboot" and related commands gained a new
1260 "--message=" option which may be used to set a free-text
1261 wall message when shutting down or rebooting the
1262 system. This message is also logged, which is useful for
1263 figuring out the reason for a reboot or shutdown a
1266 * The "systemd-resolve-host" tool's -i switch now takes
1267 network interface numbers as alternative to interface names.
1269 * A new unit file setting for services has been introduced:
1270 UtmpMode= allows configuration of how precisely systemd
1271 handles utmp and wtmp entries for the service if this is
1272 enabled. This allows writing services that appear similar to
1273 user sessions in the output of the "w", "who", "last" and
1276 * systemd-resolved will now locally synthesize DNS resource
1277 records for the "localhost" and "gateway" domains as well as
1278 the local hostname. This should ensure that clients querying
1279 RRs via resolved will get similar results as those going via
1280 NSS, if nss-myhostname is enabled.
1282 Contributions from: Alastair Hughes, Alex Crawford, Daniel
1283 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
1284 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
1285 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
1286 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
1287 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
1288 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
1289 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
1290 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
1291 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
1292 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
1293 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
1295 — Berlin, 2015-08-27
1299 * The systemd-efi-boot-generator functionality was merged into
1300 systemd-gpt-auto-generator.
1302 * systemd-networkd now supports Group Policy for vxlan
1303 devices. It can be enabled via the new boolean configuration
1304 option called 'GroupPolicyExtension='.
1306 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
1307 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
1308 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
1310 — Berlin, 2015-07-31
1314 * The python-systemd code has been removed from the systemd repository.
1315 A new repository has been created which accommodates the code from
1316 now on, and we kindly ask distributions to create a separate package
1317 for this: https://github.com/systemd/python-systemd
1319 * The systemd daemon will now reload its main configuration
1320 (/etc/systemd/system.conf) on daemon-reload.
1322 * sd-dhcp now exposes vendor specific extensions via
1323 sd_dhcp_lease_get_vendor_specific().
1325 * systemd-networkd gained a number of new configuration options.
1327 - A new boolean configuration option for TAP devices called
1328 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
1329 device, thus allowing to send and receive GSO packets.
1331 - A new tunnel configuration option called 'CopyDSCP='.
1332 If enabled, the DSCP field of ip6 tunnels is copied into the
1333 decapsulated packet.
1335 - A set of boolean bridge configuration options were added.
1336 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
1337 and 'UnicastFlood=' are now parsed by networkd and applied to the
1338 respective bridge link device via the respective IFLA_BRPORT_*
1341 - A new string configuration option to override the hostname sent
1342 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
1343 is true, networkd will use the configured hostname instead of the
1344 system hostname when sending DHCP requests.
1346 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
1347 networkd will configure the IPv6 flow-label of the tunnel device
1348 according to RFC2460.
1350 - The 'macvtap' virtual network devices are now supported, similar to
1351 the already supported 'macvlan' devices.
1353 * systemd-resolved now implements RFC5452 to improve resilience against
1354 cache poisoning. Additionally, source port randomization is enabled
1355 by default to further protect against DNS spoofing attacks.
1357 * nss-mymachines now supports translating UIDs and GIDs of running
1358 containers with user-namespaces enabled. If a container 'foo'
1359 translates a host uid 'UID' to the container uid 'TUID', then
1360 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
1361 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
1362 mapped as 'vg-foo-TGID'.
1364 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
1365 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
1366 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
1367 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
1368 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
1369 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
1370 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
1371 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
1372 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
1373 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
1375 — Berlin, 2015-07-29
1379 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
1380 There are no known issues with current sysfs, and udev does not need
1381 or should be used to work around such bugs.
1383 * udev does no longer enable USB HID power management. Several reports
1384 indicate, that some devices cannot handle that setting.
1386 * The udev accelerometer helper was removed. The functionality
1387 is now fully included in iio-sensor-proxy. But this means,
1388 older iio-sensor-proxy versions will no longer provide
1389 accelerometer/orientation data with this systemd version.
1390 Please upgrade iio-sensor-proxy to version 1.0.
1392 * networkd gained a new configuration option IPv6PrivacyExtensions=
1393 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
1394 for Stateless Address") on selected networks.
1396 * For the sake of fewer build-time dependencies and less code in the
1397 main repository, the python bindings are about to be removed in the
1398 next release. A new repository has been created which accommodates
1399 the code from now on, and we kindly ask distributions to create a
1400 separate package for this. The removal will take place in v223.
1402 https://github.com/systemd/python-systemd
1404 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
1405 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
1406 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
1407 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
1408 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
1409 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
1410 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
1411 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
1412 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
1413 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
1415 — Berlin, 2015-07-07
1419 * The sd-bus.h and sd-event.h APIs have now been declared
1420 stable and have been added to the official interface of
1421 libsystemd.so. sd-bus implements an alternative D-Bus client
1422 library, that is relatively easy to use, very efficient and
1423 supports both classic D-Bus as well as kdbus as transport
1424 backend. sd-event is a generic event loop abstraction that
1425 is built around Linux epoll, but adds features such as event
1426 prioritization or efficient timer handling. Both APIs are good
1427 choices for C programs looking for a bus and/or event loop
1428 implementation that is minimal and does not have to be
1429 portable to other kernels.
1431 * kdbus support is no longer compile-time optional. It is now
1432 always built-in. However, it can still be disabled at
1433 runtime using the kdbus=0 kernel command line setting, and
1434 that setting may be changed to default to off, by specifying
1435 --disable-kdbus at build-time. Note though that the kernel
1436 command line setting has no effect if the kdbus.ko kernel
1437 module is not installed, in which case kdbus is (obviously)
1438 also disabled. We encourage all downstream distributions to
1439 begin testing kdbus by adding it to the kernel images in the
1440 development distributions, and leaving kdbus support in
1443 * The minimal required util-linux version has been bumped to
1446 * Support for chkconfig (--enable-chkconfig) was removed in
1447 favor of calling an abstraction tool
1448 /lib/systemd/systemd-sysv-install. This needs to be
1449 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
1450 in README for details.
1452 * If there's a systemd unit and a SysV init script for the
1453 same service name, and the user executes "systemctl enable"
1454 for it (or a related call), then this will now enable both
1455 (or execute the related operation on both), not just the
1458 * The libudev API documentation has been converted from gtkdoc
1461 * gudev has been removed from the systemd tree, it is now an
1464 * The systemd-cgtop tool learnt a new --raw switch to generate
1465 "raw" (machine parsable) output.
1467 * networkd's IPForwarding= .network file setting learnt the
1468 new setting "kernel", which ensures that networkd does not
1469 change the IP forwarding sysctl from the default kernel
1472 * The systemd-logind bus API now exposes a new boolean
1473 property "Docked" that reports whether logind considers the
1474 system "docked", i.e. connected to a docking station or not.
1476 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
1477 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
1478 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
1479 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
1480 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
1481 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
1482 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
1483 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
1484 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
1485 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
1486 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
1487 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
1488 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
1489 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
1490 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
1491 Fink, Zbigniew Jędrzejewski-Szmek
1493 — Berlin, 2015-06-19
1497 * The gudev library has been extracted into a separate repository
1498 available at: https://git.gnome.org/browse/libgudev/
1499 It is now managed as part of the Gnome project. Distributions
1500 are recommended to pass --disable-gudev to systemd and use
1501 gudev from the Gnome project instead. gudev is still included
1502 in systemd, for now. It will be removed soon, though. Please
1503 also see the announcement-thread on systemd-devel:
1504 http://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
1506 * systemd now exposes a CPUUsageNSec= property for each
1507 service unit on the bus, that contains the overall consumed
1508 CPU time of a service (the sum of what each process of the
1509 service consumed). This value is only available if
1510 CPUAccounting= is turned on for a service, and is then shown
1511 in the "systemctl status" output.
1513 * Support for configuring alternative mappings of the old SysV
1514 runlevels to systemd targets has been removed. They are now
1515 hardcoded in a way that runlevels 2, 3, 4 all map to
1516 multi-user.target and 5 to graphical.target (which
1517 previously was already the default behaviour).
1519 * The auto-mounter logic gained support for mount point
1520 expiry, using a new TimeoutIdleSec= setting in .automount
1521 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
1523 * The EFI System Partition (ESP) as mounted to /boot by
1524 systemd-efi-boot-generator will now be unmounted
1525 automatically after 2 minutes of not being used. This should
1526 minimize the risk of ESP corruptions.
1528 * New /etc/fstab options x-systemd.requires= and
1529 x-systemd.requires-mounts-for= are now supported to express
1530 additional dependencies for mounts. This is useful for
1531 journalling file systems that support external journal
1532 devices or overlay file systems that require underlying file
1533 systems to be mounted.
1535 * systemd does not support direct live-upgrades (via systemctl
1536 daemon-reexec) from versions older than v44 anymore. As no
1537 distribution we are aware of shipped such old versions in a
1538 stable release this should not be problematic.
1540 * When systemd forks off a new per-connection service instance
1541 it will now set the $REMOTE_ADDR environment variable to the
1542 remote IP address, and $REMOTE_PORT environment variable to
1543 the remote IP port. This behaviour is similar to the
1544 corresponding environment variables defined by CGI.
1546 * systemd-networkd gained support for uplink failure
1547 detection. The BindCarrier= option allows binding interface
1548 configuration dynamically to the link sense of other
1549 interfaces. This is useful to achieve behaviour like in
1552 * systemd-networkd gained support for configuring the DHCP
1553 client identifier to use when requesting leases.
1555 * systemd-networkd now has a per-network UseNTP= option to
1556 configure whether NTP server information acquired via DHCP
1557 is passed on to services like systemd-timesyncd.
1559 * systemd-networkd gained support for vti6 tunnels.
1561 * Note that systemd-networkd manages the sysctl variable
1562 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
1563 it is configured for since v219. The variable controls IP
1564 forwarding, and is a per-interface alternative to the global
1565 /proc/sys/net/ipv[46]/ip_forward. This setting is
1566 configurable in the IPForward= option, which defaults to
1567 "no". This means if networkd is used for an interface it is
1568 no longer sufficient to set the global sysctl option to turn
1569 on IP forwarding! Instead, the .network file option
1570 IPForward= needs to be turned on! Note that the
1571 implementation of this behaviour was broken in v219 and has
1574 * Many bonding and vxlan options are now configurable in
1577 * systemd-nspawn gained a new --property= setting to set unit
1578 properties for the container scope. This is useful for
1579 setting resource parameters (e.g "CPUShares=500") on
1580 containers started from the command line.
1582 * systemd-nspawn gained a new --private-users= switch to make
1583 use of user namespacing available on recent Linux kernels.
1585 * systemd-nspawn may now be called as part of a shell pipeline
1586 in which case the pipes used for stdin and stdout are passed
1587 directly to the process invoked in the container, without
1588 indirection via a pseudo tty.
1590 * systemd-nspawn gained a new switch to control the UNIX
1591 signal to use when killing the init process of the container
1594 * systemd-nspawn gained a new --overlay= switch for mounting
1595 overlay file systems into the container using the new kernel
1598 * When a container image is imported via systemd-importd and
1599 the host file system is not btrfs, a loopback block device
1600 file is created in /var/lib/machines.raw with a btrfs file
1601 system inside. It is then mounted to /var/lib/machines to
1602 enable btrfs features for container management. The loopback
1603 file and btrfs file system is grown as needed when container
1604 images are imported via systemd-importd.
1606 * systemd-machined/systemd-importd gained support for btrfs
1607 quota, to enforce container disk space limits on disk. This
1608 is exposed in "machinectl set-limit".
1610 * systemd-importd now can import containers from local .tar,
1611 .raw and .qcow2 images, and export them to .tar and .raw. It
1612 can also import dkr v2 images now from the network (on top
1615 * systemd-importd gained support for verifying downloaded
1616 images with gpg2 (previously only gpg1 was supported).
1618 * systemd-machined, systemd-logind, systemd: most bus calls
1619 are now accessible to unprivileged processes via
1620 PolicyKit. Also, systemd-logind will now allow users to kill
1621 their own sessions without further privileges or
1624 * systemd-shutdownd has been removed. This service was
1625 previously responsible for implementing scheduled shutdowns
1626 as exposed in /usr/bin/shutdown's time parameter. This
1627 functionality has now been moved into systemd-logind and is
1628 accessible via a bus interface.
1630 * "systemctl reboot" gained a new switch --firmware-setup that
1631 can be used to reboot into the EFI firmware setup, if that
1632 is available. systemd-logind now exposes an API on the bus
1633 to trigger such reboots, in case graphical desktop UIs want
1634 to cover this functionality.
1636 * "systemctl enable", "systemctl disable" and "systemctl mask"
1637 now support a new "--now" switch. If specified the units
1638 that are enabled will also be started, and the ones
1639 disabled/masked also stopped.
1641 * The Gummiboot EFI boot loader tool has been merged into
1642 systemd, and renamed to "systemd-boot". The bootctl tool has been
1643 updated to support systemd-boot.
1645 * An EFI kernel stub has been added that may be used to create
1646 kernel EFI binaries that contain not only the actual kernel,
1647 but also an initrd, boot splash, command line and OS release
1648 information. This combined binary can then be signed as a
1649 single image, so that the firmware can verify it all in one
1650 step. systemd-boot has special support for EFI binaries created
1651 like this and can extract OS release information from them
1652 and show them in the boot menu. This functionality is useful
1653 to implement cryptographically verified boot schemes.
1655 * Optional support has been added to systemd-fsck to pass
1656 fsck's progress report to an AF_UNIX socket in the file
1659 * udev will no longer create device symlinks for all block
1660 devices by default. A blacklist for excluding special block
1661 devices from this logic has been turned into a whitelist
1662 that requires picking block devices explicitly that require
1665 * A new (currently still internal) API sd-device.h has been
1666 added to libsystemd. This modernized API is supposed to
1667 replace libudev eventually. In fact, already much of libudev
1668 is now just a wrapper around sd-device.h.
1670 * A new hwdb database for storing metadata about pointing
1671 stick devices has been added.
1673 * systemd-tmpfiles gained support for setting file attributes
1674 similar to the "chattr" tool with new 'h' and 'H' lines.
1676 * systemd-journald will no longer unconditionally set the
1677 btrfs NOCOW flag on new journal files. This is instead done
1678 with tmpfiles snippet using the new 'h' line type. This
1679 allows easy disabling of this logic, by masking the
1680 journal-nocow.conf tmpfiles file.
1682 * systemd-journald will now translate audit message types to
1683 human readable identifiers when writing them to the
1684 journal. This should improve readability of audit messages.
1686 * The LUKS logic gained support for the offset= and skip=
1687 options in /etc/crypttab, as previously implemented by
1690 * /usr/lib/os-release gained a new optional field VARIANT= for
1691 distributions that support multiple variants (such as a
1692 desktop edition, a server edition, ...)
1694 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
1695 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
1696 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
1697 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
1698 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
1699 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
1700 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
1701 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
1702 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
1703 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
1704 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
1705 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
1706 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
1707 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
1708 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
1709 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
1710 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
1711 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
1712 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
1713 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
1714 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
1715 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
1716 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
1717 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
1718 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
1719 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
1720 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
1722 — Berlin, 2015-05-22
1726 * Introduce a new API "sd-hwdb.h" for querying the hardware
1727 metadata database. With this minimal interface one can query
1728 and enumerate the udev hwdb, decoupled from the old libudev
1729 library. libudev's interface for this is now only a wrapper
1730 around sd-hwdb. A new tool systemd-hwdb has been added to
1731 interface with and update the database.
1733 * When any of systemd's tools copies files (for example due to
1734 tmpfiles' C lines) a btrfs reflink will attempted first,
1735 before bytewise copying is done.
1737 * systemd-nspawn gained a new --ephemeral switch. When
1738 specified a btrfs snapshot is taken of the container's root
1739 directory, and immediately removed when the container
1740 terminates again. Thus, a container can be started whose
1741 changes never alter the container's root directory, and are
1742 lost on container termination. This switch can also be used
1743 for starting a container off the root file system of the
1744 host without affecting the host OS. This switch is only
1745 available on btrfs file systems.
1747 * systemd-nspawn gained a new --template= switch. It takes the
1748 path to a container tree to use as template for the tree
1749 specified via --directory=, should that directory be
1750 missing. This allows instantiating containers dynamically,
1751 on first run. This switch is only available on btrfs file
1754 * When a .mount unit refers to a mount point on which multiple
1755 mounts are stacked, and the .mount unit is stopped all of
1756 the stacked mount points will now be unmounted until no
1757 mount point remains.
1759 * systemd now has an explicit notion of supported and
1760 unsupported unit types. Jobs enqueued for unsupported unit
1761 types will now fail with an "unsupported" error code. More
1762 specifically .swap, .automount and .device units are not
1763 supported in containers, .busname units are not supported on
1764 non-kdbus systems. .swap and .automount are also not
1765 supported if their respective kernel compile time options
1768 * machinectl gained support for two new "copy-from" and
1769 "copy-to" commands for copying files from a running
1770 container to the host or vice versa.
1772 * machinectl gained support for a new "bind" command to bind
1773 mount host directories into local containers. This is
1774 currently only supported for nspawn containers.
1776 * networkd gained support for configuring bridge forwarding
1777 database entries (fdb) from .network files.
1779 * A new tiny daemon "systemd-importd" has been added that can
1780 download container images in tar, raw, qcow2 or dkr formats,
1781 and make them available locally in /var/lib/machines, so
1782 that they can run as nspawn containers. The daemon can GPG
1783 verify the downloads (not supported for dkr, since it has no
1784 provisions for verifying downloads). It will transparently
1785 decompress bz2, xz, gzip compressed downloads if necessary,
1786 and restore sparse files on disk. The daemon uses privilege
1787 separation to ensure the actual download logic runs with
1788 fewer privileges than the daemon itself. machinectl has
1789 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
1790 make the functionality of importd available to the
1791 user. With this in place the Fedora and Ubuntu "Cloud"
1792 images can be downloaded and booted as containers unmodified
1793 (the Fedora images lack the appropriate GPG signature files
1794 currently, so they cannot be verified, but this will change
1795 soon, hopefully). Note that downloading images is currently
1796 only fully supported on btrfs.
1798 * machinectl is now able to list container images found in
1799 /var/lib/machines, along with some metadata about sizes of
1800 disk and similar. If the directory is located on btrfs and
1801 quota is enabled, this includes quota display. A new command
1802 "image-status" has been added that shows additional
1803 information about images.
1805 * machinectl is now able to clone container images
1806 efficiently, if the underlying file system (btrfs) supports
1807 it, with the new "machinectl clone" command. It also
1808 gained commands for renaming and removing images, as well as
1809 marking them read-only or read-write (supported also on
1810 legacy file systems).
1812 * networkd gained support for collecting LLDP network
1813 announcements, from hardware that supports this. This is
1814 shown in networkctl output.
1816 * systemd-run gained support for a new -t (--pty) switch for
1817 invoking a binary on a pty whose input and output is
1818 connected to the invoking terminal. This allows executing
1819 processes as system services while interactively
1820 communicating with them via the terminal. Most interestingly
1821 this is supported across container boundaries. Invoking
1822 "systemd-run -t /bin/bash" is an alternative to running a
1823 full login session, the difference being that the former
1824 will not register a session, nor go through the PAM session
1827 * tmpfiles gained support for a new "v" line type for creating
1828 btrfs subvolumes. If the underlying file system is a legacy
1829 file system, this automatically degrades to creating a
1830 normal directory. Among others /var/lib/machines is now
1831 created like this at boot, should it be missing.
1833 * The directory /var/lib/containers/ has been deprecated and
1834 been replaced by /var/lib/machines. The term "machines" has
1835 been used in the systemd context as generic term for both
1836 VMs and containers, and hence appears more appropriate for
1837 this, as the directory can also contain raw images bootable
1840 * systemd-nspawn when invoked with -M but without --directory=
1841 or --image= is now capable of searching for the container
1842 root directory, subvolume or disk image automatically, in
1843 /var/lib/machines. systemd-nspawn@.service has been updated
1844 to make use of this, thus allowing it to be used for raw
1847 * A new machines.target unit has been introduced that is
1848 supposed to group all containers/VMs invoked as services on
1849 the system. systemd-nspawn@.service has been updated to
1850 integrate with that.
1852 * machinectl gained a new "start" command, for invoking a
1853 container as a service. "machinectl start foo" is mostly
1854 equivalent to "systemctl start systemd-nspawn@foo.service",
1855 but handles escaping in a nicer way.
1857 * systemd-nspawn will now mount most of the cgroupfs tree
1858 read-only into each container, with the exception of the
1859 container's own subtree in the name=systemd hierarchy.
1861 * journald now sets the special FS_NOCOW file flag for its
1862 journal files. This should improve performance on btrfs, by
1863 avoiding heavy fragmentation when journald's write-pattern
1864 is used on COW file systems. It degrades btrfs' data
1865 integrity guarantees for the files to the same levels as for
1866 ext3/ext4 however. This should be OK though as journald does
1867 its own data integrity checks and all its objects are
1868 checksummed on disk. Also, journald should handle btrfs disk
1869 full events a lot more gracefully now, by processing SIGBUS
1870 errors, and not relying on fallocate() anymore.
1872 * When journald detects that journal files it is writing to
1873 have been deleted it will immediately start new journal
1876 * systemd now provides a way to store file descriptors
1877 per-service in PID 1.This is useful for daemons to ensure
1878 that fds they require are not lost during a daemon
1879 restart. The fds are passed to the daemon on the next
1880 invocation in the same way socket activation fds are
1881 passed. This is now used by journald to ensure that the
1882 various sockets connected to all the system's stdout/stderr
1883 are not lost when journald is restarted. File descriptors
1884 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
1885 an extension to sd_notify(). Note that a limit is enforced
1886 on the number of fds a service can store in PID 1, and it
1887 defaults to 0, so that no fds may be stored, unless this is
1888 explicitly turned on.
1890 * The default TERM variable to use for units connected to a
1891 terminal, when no other value is explicitly is set is now
1892 vt220 rather than vt102. This should be fairly safe still,
1893 but allows PgUp/PgDn work.
1895 * The /etc/crypttab option header= as known from Debian is now
1898 * "loginctl user-status" and "loginctl session-status" will
1899 now show the last 10 lines of log messages of the
1900 user/session following the status output. Similar,
1901 "machinectl status" will show the last 10 log lines
1902 associated with a virtual machine or container
1903 service. (Note that this is usually not the log messages
1904 done in the VM/container itself, but simply what the
1905 container manager logs. For nspawn this includes all console
1908 * "loginctl session-status" without further argument will now
1909 show the status of the session of the caller. Similar,
1910 "lock-session", "unlock-session", "activate",
1911 "enable-linger", "disable-linger" may now be called without
1912 session/user parameter in which case they apply to the
1913 caller's session/user.
1915 * An X11 session scriptlet is now shipped that uploads
1916 $DISPLAY and $XAUTHORITY into the environment of the systemd
1917 --user daemon if a session begins. This should improve
1918 compatibility with X11 enabled applications run as systemd
1921 * Generators are now subject to masking via /etc and /run, the
1922 same way as unit files.
1924 * networkd .network files gained support for configuring
1925 per-link IPv4/IPv6 packet forwarding as well as IPv4
1926 masquerading. This is by default turned on for veth links to
1927 containers, as registered by systemd-nspawn. This means that
1928 nspawn containers run with --network-veth will now get
1929 automatic routed access to the host's networks without any
1930 further configuration or setup, as long as networkd runs on
1933 * systemd-nspawn gained the --port= (-p) switch to expose TCP
1934 or UDP posts of a container on the host. With this in place
1935 it is possible to run containers with private veth links
1936 (--network-veth), and have their functionality exposed on
1937 the host as if their services were running directly on the
1940 * systemd-nspawn's --network-veth switch now gained a short
1941 version "-n", since with the changes above it is now truly
1942 useful out-of-the-box. The systemd-nspawn@.service has been
1943 updated to make use of it too by default.
1945 * systemd-nspawn will now maintain a per-image R/W lock, to
1946 ensure that the same image is not started more than once
1947 writable. (It's OK to run an image multiple times
1948 simultaneously in read-only mode.)
1950 * systemd-nspawn's --image= option is now capable of
1951 dissecting and booting MBR and GPT disk images that contain
1952 only a single active Linux partition. Previously it
1953 supported only GPT disk images with proper GPT type
1954 IDs. This allows running cloud images from major
1955 distributions directly with systemd-nspawn, without
1958 * In addition to collecting mouse dpi data in the udev
1959 hardware database, there's now support for collecting angle
1960 information for mouse scroll wheels. The database is
1961 supposed to guarantee similar scrolling behavior on mice
1962 that it knows about. There's also support for collecting
1963 information about Touchpad types.
1965 * udev's input_id built-in will now also collect touch screen
1966 dimension data and attach it to probed devices.
1968 * /etc/os-release gained support for a Distribution Privacy
1971 * networkd gained support for creating "ipvlan", "gretap",
1972 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
1974 * systemd-tmpfiles gained support for "a" lines for setting
1977 * systemd-nspawn will now mount /tmp in the container to
1978 tmpfs, automatically.
1980 * systemd now exposes the memory.usage_in_bytes cgroup
1981 attribute and shows it for each service in the "systemctl
1982 status" output, if available.
1984 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
1985 immediate reboot is triggered. This useful if shutdown is
1986 hung and is unable to complete, to expedite the
1987 operation. Note that this kind of reboot will still unmount
1988 all file systems, and hence should not result in fsck being
1991 * A .device unit for an optical block device will now be
1992 considered active only when a medium is in the drive. Also,
1993 mount units are now bound to their backing devices thus
1994 triggering automatic unmounting when devices become
1995 unavailable. With this in place systemd will now
1996 automatically unmount left-over mounts when a CD-ROM is
1997 ejected or an USB stick is yanked from the system.
1999 * networkd-wait-online now has support for waiting for
2000 specific interfaces only (with globbing), and for giving up
2001 after a configurable timeout.
2003 * networkd now exits when idle. It will be automatically
2004 restarted as soon as interfaces show up, are removed or
2005 change state. networkd will stay around as long as there is
2006 at least one DHCP state machine or similar around, that keep
2009 * networkd may now configure IPv6 link-local addressing in
2010 addition to IPv4 link-local addressing.
2012 * The IPv6 "token" for use in SLAAC may now be configured for
2013 each .network interface in networkd.
2015 * Routes configured with networkd may now be assigned a scope
2018 * networkd's [Match] sections now support globbing and lists
2019 of multiple space-separated matches per item.
2021 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
2022 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
2023 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
2024 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
2025 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
2026 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
2027 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
2028 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
2029 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
2030 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
2031 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
2032 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
2033 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
2034 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
2035 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
2036 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
2037 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
2038 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
2039 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
2040 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
2041 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
2042 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
2043 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
2044 Hoffmann, Zbigniew Jędrzejewski-Szmek
2046 — Berlin, 2015-02-16
2050 * When querying unit file enablement status (for example via
2051 "systemctl is-enabled"), a new state "indirect" is now known
2052 which indicates that a unit might not be enabled itself, but
2053 another unit listed in its Also= setting might be.
2055 * Similar to the various existing ConditionXYZ= settings for
2056 units, there are now matching AssertXYZ= settings. While
2057 failing conditions cause a unit to be skipped, but its job
2058 to succeed, failing assertions declared like this will cause
2059 a unit start operation and its job to fail.
2061 * hostnamed now knows a new chassis type "embedded".
2063 * systemctl gained a new "edit" command. When used on a unit
2064 file, this allows extending unit files with .d/ drop-in
2065 configuration snippets or editing the full file (after
2066 copying it from /usr/lib to /etc). This will invoke the
2067 user's editor (as configured with $EDITOR), and reload the
2068 modified configuration after editing.
2070 * "systemctl status" now shows the suggested enablement state
2071 for a unit, as declared in the (usually vendor-supplied)
2072 system preset files.
2074 * nss-myhostname will now resolve the single-label host name
2075 "gateway" to the locally configured default IP routing
2076 gateways, ordered by their metrics. This assigns a stable
2077 name to the used gateways, regardless which ones are
2078 currently configured. Note that the name will only be
2079 resolved after all other name sources (if nss-myhostname is
2080 configured properly) and should hence not negatively impact
2081 systems that use the single-label host name "gateway" in
2084 * systemd-inhibit now allows filtering by mode when listing
2087 * Scope and service units gained a new "Delegate" boolean
2088 property, which, when set, allows processes running inside the
2089 unit to further partition resources. This is primarily
2090 useful for systemd user instances as well as container
2093 * journald will now pick up audit messages directly from
2094 the kernel, and log them like any other log message. The
2095 audit fields are split up and fully indexed. This means that
2096 journalctl in many ways is now a (nicer!) alternative to
2097 ausearch, the traditional audit client. Note that this
2098 implements only a minimal audit client. If you want the
2099 special audit modes like reboot-on-log-overflow, please use
2100 the traditional auditd instead, which can be used in
2101 parallel to journald.
2103 * The ConditionSecurity= unit file option now understands the
2104 special string "audit" to check whether auditing is
2107 * journalctl gained two new commands --vacuum-size= and
2108 --vacuum-time= to delete old journal files until the
2109 remaining ones take up no more than the specified size on disk,
2110 or are not older than the specified time.
2112 * A new, native PPPoE library has been added to sd-network,
2113 systemd's library of light-weight networking protocols. This
2114 library will be used in a future version of networkd to
2115 enable PPPoE communication without an external pppd daemon.
2117 * The busctl tool now understands a new "capture" verb that
2118 works similar to "monitor", but writes a packet capture
2119 trace to STDOUT that can be redirected to a file which is
2120 compatible with libcap's capture file format. This can then
2121 be loaded in Wireshark and similar tools to inspect bus
2124 * The busctl tool now understands a new "tree" verb that shows
2125 the object trees of a specific service on the bus, or of all
2128 * The busctl tool now understands a new "introspect" verb that
2129 shows all interfaces and members of objects on the bus,
2130 including their signature and values. This is particularly
2131 useful to get more information about bus objects shown by
2132 the new "busctl tree" command.
2134 * The busctl tool now understands new verbs "call",
2135 "set-property" and "get-property" for invoking bus method
2136 calls, setting and getting bus object properties in a
2139 * busctl gained a new --augment-creds= argument that controls
2140 whether the tool shall augment credential information it
2141 gets from the bus with data from /proc, in a possibly
2144 * nspawn's --link-journal= switch gained two new values
2145 "try-guest" and "try-host" that work like "guest" and
2146 "host", but do not fail if the host has no persistent
2147 journalling enabled. -j is now equivalent to
2148 --link-journal=try-guest.
2150 * macvlan network devices created by nspawn will now have
2151 stable MAC addresses.
2153 * A new SmackProcessLabel= unit setting has been added, which
2154 controls the SMACK security label processes forked off by
2155 the respective unit shall use.
2157 * If compiled with --enable-xkbcommon, systemd-localed will
2158 verify x11 keymap settings by compiling the given keymap. It
2159 will spew out warnings if the compilation fails. This
2160 requires libxkbcommon to be installed.
2162 * When a coredump is collected, a larger number of metadata
2163 fields is now collected and included in the journal records
2164 created for it. More specifically, control group membership,
2165 environment variables, memory maps, working directory,
2166 chroot directory, /proc/$PID/status, and a list of open file
2167 descriptors is now stored in the log entry.
2169 * The udev hwdb now contains DPI information for mice. For
2172 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
2174 * All systemd programs that read standalone configuration
2175 files in /etc now also support a corresponding series of
2176 .conf.d configuration directories in /etc/, /run/,
2177 /usr/local/lib/, /usr/lib/, and (if configured with
2178 --enable-split-usr) /lib/. In particular, the following
2179 configuration files now have corresponding configuration
2180 directories: system.conf user.conf, logind.conf,
2181 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
2182 resolved.conf, timesyncd.conf, journal-remote.conf, and
2183 journal-upload.conf. Note that distributions should use the
2184 configuration directories in /usr/lib/; the directories in
2185 /etc/ are reserved for the system administrator.
2187 * systemd-rfkill will no longer take the rfkill device name
2188 into account when storing rfkill state on disk, as the name
2189 might be dynamically assigned and not stable. Instead, the
2190 ID_PATH udev variable combined with the rfkill type (wlan,
2191 bluetooth, ...) is used.
2193 * A new service systemd-machine-id-commit.service has been
2194 added. When used on systems where /etc is read-only during
2195 boot, and /etc/machine-id is not initialized (but an empty
2196 file), this service will copy the temporary machine ID
2197 created as replacement into /etc after the system is fully
2198 booted up. This is useful for systems that are freshly
2199 installed with a non-initialized machine ID, but should get
2200 a fixed machine ID for subsequent boots.
2202 * networkd's .netdev files now provide a large set of
2203 configuration parameters for VXLAN devices. Similarly, the
2204 bridge port cost parameter is now configurable in .network
2205 files. There's also new support for configuring IP source
2206 routing. networkd .link files gained support for a new
2207 OriginalName= match that is useful to match against the
2208 original interface name the kernel assigned. .network files
2209 may include MTU= and MACAddress= fields for altering the MTU
2210 and MAC address while being connected to a specific network
2213 * The LUKS logic gained supported for configuring
2214 UUID-specific key files. There's also new support for naming
2215 LUKS device from the kernel command line, using the new
2216 luks.name= argument.
2218 * Timer units may now be transiently created via the bus API
2219 (this was previously already available for scope and service
2220 units). In addition it is now possible to create multiple
2221 transient units at the same time with a single bus call. The
2222 "systemd-run" tool has been updated to make use of this for
2223 running commands on a specified time, in at(1)-style.
2225 * tmpfiles gained support for "t" lines, for assigning
2226 extended attributes to files. Among other uses this may be
2227 used to assign SMACK labels to files.
2229 Contributions from: Alin Rauta, Alison Chaiken, Andrej
2230 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
2231 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
2232 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
2233 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
2234 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
2235 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
2236 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
2237 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
2238 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
2239 Michael Biebl, Michael Chapman, Michael Marineau, Michal
2240 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
2241 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
2242 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
2243 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
2244 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2245 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
2246 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
2248 — Berlin, 2014-12-10
2252 * journalctl gained the new options -t/--identifier= to match
2253 on the syslog identifier (aka "tag"), as well as --utc to
2254 show log timestamps in the UTC timezone. journalctl now also
2255 accepts -n/--lines=all to disable line capping in a pager.
2257 * journalctl gained a new switch, --flush, that synchronously
2258 flushes logs from /run/log/journal to /var/log/journal if
2259 persistent storage is enabled. systemd-journal-flush.service
2260 now waits until the operation is complete.
2262 * Services can notify the manager before they start a reload
2263 (by sending RELOADING=1) or shutdown (by sending
2264 STOPPING=1). This allows the manager to track and show the
2265 internal state of daemons and closes a race condition when
2266 the process is still running but has closed its D-Bus
2269 * Services with Type=oneshot do not have to have any ExecStart
2272 * User units are now loaded also from
2273 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
2274 /run/systemd/user directory that was already previously
2275 supported, but is under the control of the user.
2277 * Job timeouts (i.e. time-outs on the time a job that is
2278 queued stays in the run queue) can now optionally result in
2279 immediate reboot or power-off actions (JobTimeoutAction= and
2280 JobTimeoutRebootArgument=). This is useful on ".target"
2281 units, to limit the maximum time a target remains
2282 undispatched in the run queue, and to trigger an emergency
2283 operation in such a case. This is now used by default to
2284 turn off the system if boot-up (as defined by everything in
2285 basic.target) hangs and does not complete for at least
2286 15min. Also, if power-off or reboot hang for at least 30min
2287 an immediate power-off/reboot operation is triggered. This
2288 functionality is particularly useful to increase reliability
2289 on embedded devices, but also on laptops which might
2290 accidentally get powered on when carried in a backpack and
2291 whose boot stays stuck in a hard disk encryption passphrase
2294 * systemd-logind can be configured to also handle lid switch
2295 events even when the machine is docked or multiple displays
2296 are attached (HandleLidSwitchDocked= option).
2298 * A helper binary and a service have been added which can be
2299 used to resume from hibernation in the initramfs. A
2300 generator will parse the resume= option on the kernel
2301 command line to trigger resume.
2303 * A user console daemon systemd-consoled has been
2304 added. Currently, it is a preview, and will so far open a
2305 single terminal on each session of the user marked as
2306 Desktop=systemd-console.
2308 * Route metrics can be specified for DHCP routes added by
2311 * The SELinux context of socket-activated services can be set
2312 from the information provided by the networking stack
2313 (SELinuxContextFromNet= option).
2315 * Userspace firmware loading support has been removed and
2316 the minimum supported kernel version is thus bumped to 3.7.
2318 * Timeout for udev workers has been increased from 1 to 3
2319 minutes, but a warning will be printed after 1 minute to
2320 help diagnose kernel modules that take a long time to load.
2322 * Udev rules can now remove tags on devices with TAG-="foobar".
2324 * systemd's readahead implementation has been removed. In many
2325 circumstances it didn't give expected benefits even for
2326 rotational disk drives and was becoming less relevant in the
2327 age of SSDs. As none of the developers has been using
2328 rotating media anymore, and nobody stepped up to actively
2329 maintain this component of systemd it has now been removed.
2331 * Swap units can use Options= to specify discard options.
2332 Discard options specified for swaps in /etc/fstab are now
2335 * Docker containers are now detected as a separate type of
2338 * The Password Agent protocol gained support for queries where
2339 the user input is shown, useful e.g. for user names.
2340 systemd-ask-password gained a new --echo option to turn that
2343 * The default sysctl.d/ snippets will now set:
2345 net.core.default_qdisc = fq_codel
2347 This selects Fair Queuing Controlled Delay as the default
2348 queuing discipline for network interfaces. fq_codel helps
2349 fight the network bufferbloat problem. It is believed to be
2350 a good default with no tuning required for most workloads.
2351 Downstream distributions may override this choice. On 10Gbit
2352 servers that do not do forwarding, "fq" may perform better.
2353 Systems without a good clocksource should use "pfifo_fast".
2355 * If kdbus is enabled during build a new option BusPolicy= is
2356 available for service units, that allows locking all service
2357 processes into a stricter bus policy, in order to limit
2358 access to various bus services, or even hide most of them
2359 from the service's view entirely.
2361 * networkctl will now show the .network and .link file
2362 networkd has applied to a specific interface.
2364 * sd-login gained a new API call sd_session_get_desktop() to
2365 query which desktop environment has been selected for a
2368 * UNIX utmp support is now compile-time optional to support
2369 legacy-free systems.
2371 * systemctl gained two new commands "add-wants" and
2372 "add-requires" for pulling in units from specific targets
2375 * If the word "rescue" is specified on the kernel command line
2376 the system will now boot into rescue mode (aka
2377 rescue.target), which was previously available only by
2378 specifying "1" or "systemd.unit=rescue.target" on the kernel
2379 command line. This new kernel command line option nicely
2380 mirrors the already existing "emergency" kernel command line
2383 * New kernel command line options mount.usr=, mount.usrflags=,
2384 mount.usrfstype= have been added that match root=, rootflags=,
2385 rootfstype= but allow mounting a specific file system to
2388 * The $NOTIFY_SOCKET is now also passed to control processes of
2389 services, not only the main process.
2391 * This version reenables support for fsck's -l switch. This
2392 means at least version v2.25 of util-linux is required for
2393 operation, otherwise dead-locks on device nodes may
2394 occur. Again: you need to update util-linux to at least
2395 v2.25 when updating systemd to v217.
2397 * The "multi-seat-x" tool has been removed from systemd, as
2398 its functionality has been integrated into X servers 1.16,
2399 and the tool is hence redundant. It is recommended to update
2400 display managers invoking this tool to simply invoke X
2401 directly from now on, again.
2403 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
2404 message flag has been added for all of systemd's PolicyKit
2405 authenticated method calls has been added. In particular
2406 this now allows optional interactive authorization via
2407 PolicyKit for many of PID1's privileged operations such as
2408 unit file enabling and disabling.
2410 * "udevadm hwdb --update" learnt a new switch "--usr" for
2411 placing the rebuilt hardware database in /usr instead of
2412 /etc. When used only hardware database entries stored in
2413 /usr will be used, and any user database entries in /etc are
2414 ignored. This functionality is useful for vendors to ship a
2415 pre-built database on systems where local configuration is
2416 unnecessary or unlikely.
2418 * Calendar time specifications in .timer units now also
2419 understand the strings "semi-annually", "quarterly" and
2420 "minutely" as shortcuts (in addition to the preexisting
2421 "anually", "hourly", ...).
2423 * systemd-tmpfiles will now correctly create files in /dev
2424 at boot which are marked for creation only at boot. It is
2425 recommended to always create static device nodes with 'c!'
2426 and 'b!', so that they are created only at boot and not
2427 overwritten at runtime.
2429 * When the watchdog logic is used for a service (WatchdogSec=)
2430 and the watchdog timeout is hit the service will now be
2431 terminated with SIGABRT (instead of just SIGTERM), in order
2432 to make sure a proper coredump and backtrace is
2433 generated. This ensures that hanging services will result in
2434 similar coredump/backtrace behaviour as services that hit a
2437 Contributions from: Andreas Henriksson, Andrei Borzenkov,
2438 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
2439 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
2440 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
2441 Herrmann, David Sommerseth, David Strauss, Emil Renner
2442 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
2443 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
2444 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
2445 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
2446 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
2447 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
2448 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
2449 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
2450 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
2451 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
2452 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
2453 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
2454 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
2455 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
2456 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
2457 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
2460 — Berlin, 2014-10-28
2464 * timedated no longer reads NTP implementation unit names from
2465 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
2466 implementations should add a
2468 Conflicts=systemd-timesyncd.service
2470 to their unit files to take over and replace systemd's NTP
2471 default functionality.
2473 * systemd-sysusers gained a new line type "r" for configuring
2474 which UID/GID ranges to allocate system users/groups
2475 from. Lines of type "u" may now add an additional column
2476 that specifies the home directory for the system user to be
2477 created. Also, systemd-sysusers may now optionally read user
2478 information from STDIN instead of a file. This is useful for
2479 invoking it from RPM preinst scriptlets that need to create
2480 users before the first RPM file is installed since these
2481 files might need to be owned by them. A new
2482 %sysusers_create_inline RPM macro has been introduced to do
2483 just that. systemd-sysusers now updates the shadow files as
2484 well as the user/group databases, which should enhance
2485 compatibility with certain tools like grpck.
2487 * A number of bus APIs of PID 1 now optionally consult
2488 PolicyKit to permit access for otherwise unprivileged
2489 clients under certain conditions. Note that this currently
2490 doesn't support interactive authentication yet, but this is
2491 expected to be added eventually, too.
2493 * /etc/machine-info now has new fields for configuring the
2494 deployment environment of the machine, as well as the
2495 location of the machine. hostnamectl has been updated with
2496 new command to update these fields.
2498 * systemd-timesyncd has been updated to automatically acquire
2499 NTP server information from systemd-networkd, which might
2500 have been discovered via DHCP.
2502 * systemd-resolved now includes a caching DNS stub resolver
2503 and a complete LLMNR name resolution implementation. A new
2504 NSS module "nss-resolve" has been added which can be used
2505 instead of glibc's own "nss-dns" to resolve hostnames via
2506 systemd-resolved. Hostnames, addresses and arbitrary RRs may
2507 be resolved via systemd-resolved D-Bus APIs. In contrast to
2508 the glibc internal resolver systemd-resolved is aware of
2509 multi-homed system, and keeps DNS server and caches separate
2510 and per-interface. Queries are sent simultaneously on all
2511 interfaces that have DNS servers configured, in order to
2512 properly handle VPNs and local LANs which might resolve
2513 separate sets of domain names. systemd-resolved may acquire
2514 DNS server information from systemd-networkd automatically,
2515 which in turn might have discovered them via DHCP. A tool
2516 "systemd-resolve-host" has been added that may be used to
2517 query the DNS logic in resolved. systemd-resolved implements
2518 IDNA and automatically uses IDNA or UTF-8 encoding depending
2519 on whether classic DNS or LLMNR is used as transport. In the
2520 next releases we intend to add a DNSSEC and mDNS/DNS-SD
2521 implementation to systemd-resolved.
2523 * A new NSS module nss-mymachines has been added, that
2524 automatically resolves the names of all local registered
2525 containers to their respective IP addresses.
2527 * A new client tool "networkctl" for systemd-networkd has been
2528 added. It currently is entirely passive and will query
2529 networking configuration from udev, rtnetlink and networkd,
2530 and present it to the user in a very friendly
2531 way. Eventually, we hope to extend it to become a full
2532 control utility for networkd.
2534 * .socket units gained a new DeferAcceptSec= setting that
2535 controls the kernels' TCP_DEFER_ACCEPT sockopt for
2536 TCP. Similarly, support for controlling TCP keep-alive
2537 settings has been added (KeepAliveTimeSec=,
2538 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
2539 turning off Nagle's algorithm on TCP has been added
2542 * logind learned a new session type "web", for use in projects
2543 like Cockpit which register web clients as PAM sessions.
2545 * timer units with at least one OnCalendar= setting will now
2546 be started only after timer-sync.target has been
2547 reached. This way they will not elapse before the system
2548 clock has been corrected by a local NTP client or
2549 similar. This is particular useful on RTC-less embedded
2550 machines, that come up with an invalid system clock.
2552 * systemd-nspawn's --network-veth= switch should now result in
2553 stable MAC addresses for both the outer and the inner side
2556 * systemd-nspawn gained a new --volatile= switch for running
2557 container instances with /etc or /var unpopulated.
2559 * The kdbus client code has been updated to use the new Linux
2560 3.17 memfd subsystem instead of the old kdbus-specific one.
2562 * systemd-networkd's DHCP client and server now support
2563 FORCERENEW. There are also new configuration options to
2564 configure the vendor client identifier and broadcast mode
2567 * systemd will no longer inform the kernel about the current
2568 timezone, as this is necessarily incorrect and racy as the
2569 kernel has no understanding of DST and similar
2570 concepts. This hence means FAT timestamps will be always
2571 considered UTC, similar to what Android is already
2572 doing. Also, when the RTC is configured to the local time
2573 (rather than UTC) systemd will never synchronize back to it,
2574 as this might confuse Windows at a later boot.
2576 * systemd-analyze gained a new command "verify" for offline
2577 validation of unit files.
2579 * systemd-networkd gained support for a couple of additional
2580 settings for bonding networking setups. Also, the metric for
2581 statically configured routes may now be configured. For
2582 network interfaces where this is appropriate the peer IP
2583 address may now be configured.
2585 * systemd-networkd's DHCP client will no longer request
2586 broadcasting by default, as this tripped up some networks.
2587 For hardware where broadcast is required the feature should
2588 be switched back on using RequestBroadcast=yes.
2590 * systemd-networkd will now set up IPv4LL addresses (when
2591 enabled) even if DHCP is configured successfully.
2593 * udev will now default to respect network device names given
2594 by the kernel when the kernel indicates that these are
2595 predictable. This behavior can be tweaked by changing
2596 NamePolicy= in the relevant .link file.
2598 * A new library systemd-terminal has been added that
2599 implements full TTY stream parsing and rendering. This
2600 library is supposed to be used later on for implementing a
2601 full userspace VT subsystem, replacing the current kernel
2604 * A new tool systemd-journal-upload has been added to push
2605 journal data to a remote system running
2606 systemd-journal-remote.
2608 * journald will no longer forward all local data to another
2609 running syslog daemon. This change has been made because
2610 rsyslog (which appears to be the most commonly used syslog
2611 implementation these days) no longer makes use of this, and
2612 instead pulls the data out of the journal on its own. Since
2613 forwarding the messages to a non-existent syslog server is
2614 more expensive than we assumed we have now turned this
2615 off. If you run a syslog server that is not a recent rsyslog
2616 version, you have to turn this option on again
2617 (ForwardToSyslog= in journald.conf).
2619 * journald now optionally supports the LZ4 compressor for
2620 larger journal fields. This compressor should perform much
2621 better than XZ which was the previous default.
2623 * machinectl now shows the IP addresses of local containers,
2624 if it knows them, plus the interface name of the container.
2626 * A new tool "systemd-escape" has been added that makes it
2627 easy to escape strings to build unit names and similar.
2629 * sd_notify() messages may now include a new ERRNO= field
2630 which is parsed and collected by systemd and shown among the
2631 "systemctl status" output for a service.
2633 * A new component "systemd-firstboot" has been added that
2634 queries the most basic systemd information (timezone,
2635 hostname, root password) interactively on first
2636 boot. Alternatively it may also be used to provision these
2637 things offline on OS images installed into directories.
2639 * The default sysctl.d/ snippets will now set
2641 net.ipv4.conf.default.promote_secondaries=1
2643 This has the benefit of no flushing secondary IP addresses
2644 when primary addresses are removed.
2646 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
2647 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
2648 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
2649 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
2650 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
2651 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
2652 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
2653 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
2654 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
2655 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
2656 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
2657 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
2658 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
2659 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
2660 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
2662 — Berlin, 2014-08-19
2666 * A new tool systemd-sysusers has been added. This tool
2667 creates system users and groups in /etc/passwd and
2668 /etc/group, based on static declarative system user/group
2669 definitions in /usr/lib/sysusers.d/. This is useful to
2670 enable factory resets and volatile systems that boot up with
2671 an empty /etc directory, and thus need system users and
2672 groups created during early boot. systemd now also ships
2673 with two default sysusers.d/ files for the most basic
2674 users and groups systemd and the core operating system
2677 * A new tmpfiles snippet has been added that rebuilds the
2678 essential files in /etc on boot, should they be missing.
2680 * A directive for ensuring automatic clean-up of
2681 /var/cache/man/ has been removed from the default
2682 configuration. This line should now be shipped by the man
2683 implementation. The necessary change has been made to the
2684 man-db implementation. Note that you need to update your man
2685 implementation to one that ships this line, otherwise no
2686 automatic clean-up of /var/cache/man will take place.
2688 * A new condition ConditionNeedsUpdate= has been added that
2689 may conditionalize services to only run when /etc or /var
2690 are "older" than the vendor operating system resources in
2691 /usr. This is useful for reconstructing or updating /etc
2692 after an offline update of /usr or a factory reset, on the
2693 next reboot. Services that want to run once after such an
2694 update or reset should use this condition and order
2695 themselves before the new systemd-update-done.service, which
2696 will mark the two directories as fully updated. A number of
2697 service files have been added making use of this, to rebuild
2698 the udev hardware database, the journald message catalog and
2699 dynamic loader cache (ldconfig). The systemd-sysusers tool
2700 described above also makes use of this now. With this in
2701 place it is now possible to start up a minimal operating
2702 system with /etc empty cleanly. For more information on the
2703 concepts involved see this recent blog story:
2705 http://0pointer.de/blog/projects/stateless.html
2707 * A new system group "input" has been introduced, and all
2708 input device nodes get this group assigned. This is useful
2709 for system-level software to get access to input devices. It
2710 complements what is already done for "audio" and "video".
2712 * systemd-networkd learnt minimal DHCPv4 server support in
2713 addition to the existing DHCPv4 client support. It also
2714 learnt DHCPv6 client and IPv6 Router Solicitation client
2715 support. The DHCPv4 client gained support for static routes
2716 passed in from the server. Note that the [DHCPv4] section
2717 known in older systemd-networkd versions has been renamed to
2718 [DHCP] and is now also used by the DHCPv6 client. Existing
2719 .network files using settings of this section should be
2720 updated, though compatibility is maintained. Optionally, the
2721 client hostname may now be sent to the DHCP server.
2723 * networkd gained support for vxlan virtual networks as well
2724 as tun/tap and dummy devices.
2726 * networkd gained support for automatic allocation of address
2727 ranges for interfaces from a system-wide pool of
2728 addresses. This is useful for dynamically managing a large
2729 number of interfaces with a single network configuration
2730 file. In particular this is useful to easily assign
2731 appropriate IP addresses to the veth links of a large number
2732 of nspawn instances.
2734 * RPM macros for processing sysusers, sysctl and binfmt
2735 drop-in snippets at package installation time have been
2738 * The /etc/os-release file should now be placed in
2739 /usr/lib/os-release. The old location is automatically
2740 created as symlink. /usr/lib is the more appropriate
2741 location of this file, since it shall actually describe the
2742 vendor operating system shipped in /usr, and not the
2743 configuration stored in /etc.
2745 * .mount units gained a new boolean SloppyOptions= setting
2746 that maps to mount(8)'s -s option which enables permissive
2747 parsing of unknown mount options.
2749 * tmpfiles learnt a new "L+" directive which creates a symlink
2750 but (unlike "L") deletes a pre-existing file first, should
2751 it already exist and not already be the correct
2752 symlink. Similarly, "b+", "c+" and "p+" directives have been
2753 added as well, which create block and character devices, as
2754 well as fifos in the filesystem, possibly removing any
2755 pre-existing files of different types.
2757 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
2758 'argument' field (which so far specified the source to
2759 symlink/copy the files from) is now optional. If omitted the
2760 same file os copied from /usr/share/factory/ suffixed by the
2761 full destination path. This is useful for populating /etc
2762 with essential files, by copying them from vendor defaults
2763 shipped in /usr/share/factory/etc.
2765 * A new command "systemctl preset-all" has been added that
2766 applies the service preset settings to all installed unit
2767 files. A new switch --preset-mode= has been added that
2768 controls whether only enable or only disable operations
2771 * A new command "systemctl is-system-running" has been added
2772 that allows checking the overall state of the system, for
2773 example whether it is fully up and running.
2775 * When the system boots up with an empty /etc, the equivalent
2776 to "systemctl preset-all" is executed during early boot, to
2777 make sure all default services are enabled after a factory
2780 * systemd now contains a minimal preset file that enables the
2781 most basic services systemd ships by default.
2783 * Unit files' [Install] section gained a new DefaultInstance=
2784 field for defining the default instance to create if a
2785 template unit is enabled with no instance specified.
2787 * A new passive target cryptsetup-pre.target has been added
2788 that may be used by services that need to make they run and
2789 finish before the first LUKS cryptographic device is set up.
2791 * The /dev/loop-control and /dev/btrfs-control device nodes
2792 are now owned by the "disk" group by default, opening up
2793 access to this group.
2795 * systemd-coredump will now automatically generate a
2796 stack trace of all core dumps taking place on the system,
2797 based on elfutils' libdw library. This stack trace is logged
2800 * systemd-coredump may now optionally store coredumps directly
2801 on disk (in /var/lib/systemd/coredump, possibly compressed),
2802 instead of storing them unconditionally in the journal. This
2803 mode is the new default. A new configuration file
2804 /etc/systemd/coredump.conf has been added to configure this
2805 and other parameters of systemd-coredump.
2807 * coredumpctl gained a new "info" verb to show details about a
2808 specific coredump. A new switch "-1" has also been added
2809 that makes sure to only show information about the most
2810 recent entry instead of all entries. Also, as the tool is
2811 generally useful now the "systemd-" prefix of the binary
2812 name has been removed. Distributions that want to maintain
2813 compatibility with the old name should add a symlink from
2814 the old name to the new name.
2816 * journald's SplitMode= now defaults to "uid". This makes sure
2817 that unprivileged users can access their own coredumps with
2818 coredumpctl without restrictions.
2820 * New kernel command line options "systemd.wants=" (for
2821 pulling an additional unit during boot), "systemd.mask="
2822 (for masking a specific unit for the boot), and
2823 "systemd.debug-shell" (for enabling the debug shell on tty9)
2824 have been added. This is implemented in the new generator
2825 "systemd-debug-generator".
2827 * systemd-nspawn will now by default filter a couple of
2828 syscalls for containers, among them those required for
2829 kernel module loading, direct x86 IO port access, swap
2830 management, and kexec. Most importantly though
2831 open_by_handle_at() is now prohibited for containers,
2832 closing a hole similar to a recently discussed vulnerability
2833 in docker regarding access to files on file hierarchies the
2834 container should normally not have access to. Note that, for
2835 nspawn, we generally make no security claims anyway (and
2836 this is explicitly documented in the man page), so this is
2837 just a fix for one of the most obvious problems.
2839 * A new man page file-hierarchy(7) has been added that
2840 contains a minimized, modernized version of the file system
2841 layout systemd expects, similar in style to the FHS
2842 specification or hier(5). A new tool systemd-path(1) has
2843 been added to query many of these paths for the local
2846 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
2847 longer done. Since the directory now has a per-user size
2848 limit, and is cleaned on logout this appears unnecessary,
2849 in particular since this now brings the lifecycle of this
2850 directory closer in line with how IPC objects are handled.
2852 * systemd.pc now exports a number of additional directories,
2853 including $libdir (which is useful to identify the library
2854 path for the primary architecture of the system), and a
2855 couple of drop-in directories.
2857 * udev's predictable network interface names now use the dev_port
2858 sysfs attribute, introduced in linux 3.15 instead of dev_id to
2859 distinguish between ports of the same PCI function. dev_id should
2860 only be used for ports using the same HW address, hence the need
2863 * machined has been updated to export the OS version of a
2864 container (read from /etc/os-release and
2865 /usr/lib/os-release) on the bus. This is now shown in
2866 "machinectl status" for a machine.
2868 * A new service setting RestartForceExitStatus= has been
2869 added. If configured to a set of exit signals or process
2870 return values, the service will be restarted when the main
2871 daemon process exits with any of them, regardless of the
2874 * systemctl's -H switch for connecting to remote systemd
2875 machines has been extended so that it may be used to
2876 directly connect to a specific container on the
2877 host. "systemctl -H root@foobar:waldi" will now connect as
2878 user "root" to host "foobar", and then proceed directly to
2879 the container named "waldi". Note that currently you have to
2880 authenticate as user "root" for this to work, as entering
2881 containers is a privileged operation.
2883 Contributions from: Andreas Henriksson, Benjamin Steinwender,
2884 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
2885 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
2886 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
2887 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
2888 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
2889 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
2890 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
2891 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
2892 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
2893 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
2894 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
2896 — Berlin, 2014-07-03
2900 * As an experimental feature, udev now tries to lock the
2901 disk device node (flock(LOCK_SH|LOCK_NB)) while it
2902 executes events for the disk or any of its partitions.
2903 Applications like partitioning programs can lock the
2904 disk device node (flock(LOCK_EX)) and claim temporary
2905 device ownership that way; udev will entirely skip all event
2906 handling for this disk and its partitions. If the disk
2907 was opened for writing, the close will trigger a partition
2908 table rescan in udev's "watch" facility, and if needed
2909 synthesize "change" events for the disk and all its partitions.
2910 This is now unconditionally enabled, and if it turns out to
2911 cause major problems, we might turn it on only for specific
2912 devices, or might need to disable it entirely. Device Mapper
2913 devices are excluded from this logic.
2915 * We temporarily dropped the "-l" switch for fsck invocations,
2916 since they collide with the flock() logic above. util-linux
2917 upstream has been changed already to avoid this conflict,
2918 and we will readd "-l" as soon as util-linux with this
2919 change has been released.
2921 * The dependency on libattr has been removed. Since a long
2922 time, the extended attribute calls have moved to glibc, and
2923 libattr is thus unnecessary.
2925 * Virtualization detection works without priviliges now. This
2926 means the systemd-detect-virt binary no longer requires
2927 CAP_SYS_PTRACE file capabilities, and our daemons can run
2928 with fewer privileges.
2930 * systemd-networkd now runs under its own "systemd-network"
2931 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
2932 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
2933 loses the ability to write to files owned by root this way.
2935 * Similarly, systemd-resolved now runs under its own
2936 "systemd-resolve" user with no capabilities remaining.
2938 * Similarly, systemd-bus-proxyd now runs under its own
2939 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
2941 * systemd-networkd gained support for setting up "veth"
2942 virtual Ethernet devices for container connectivity, as well
2943 as GRE and VTI tunnels.
2945 * systemd-networkd will no longer automatically attempt to
2946 manually load kernel modules necessary for certain tunnel
2947 transports. Instead, it is assumed the kernel loads them
2948 automatically when required. This only works correctly on
2949 very new kernels. On older kernels, please consider adding
2950 the kernel modules to /etc/modules-load.d/ as a work-around.
2952 * The resolv.conf file systemd-resolved generates has been
2953 moved to /run/systemd/resolve/. If you have a symlink from
2954 /etc/resolv.conf, it might be necessary to correct it.
2956 * Two new service settings, ProtectHome= and ProtectSystem=,
2957 have been added. When enabled, they will make the user data
2958 (such as /home) inaccessible or read-only and the system
2959 (such as /usr) read-only, for specific services. This allows
2960 very light-weight per-service sandboxing to avoid
2961 modifications of user data or system files from
2962 services. These two new switches have been enabled for all
2963 of systemd's long-running services, where appropriate.
2965 * Socket units gained new SocketUser= and SocketGroup=
2966 settings to set the owner user and group of AF_UNIX sockets
2967 and FIFOs in the file system.
2969 * Socket units gained a new RemoveOnStop= setting. If enabled,
2970 all FIFOS and sockets in the file system will be removed
2971 when the specific socket unit is stopped.
2973 * Socket units gained a new Symlinks= setting. It takes a list
2974 of symlinks to create to file system sockets or FIFOs
2975 created by the specific Unix sockets. This is useful to
2976 manage symlinks to socket nodes with the same life-cycle as
2979 * The /dev/log socket and /dev/initctl FIFO have been moved to
2980 /run, and have been replaced by symlinks. This allows
2981 connecting to these facilities even if PrivateDevices=yes is
2982 used for a service (which makes /dev/log itself unavailable,
2983 but /run is left). This also has the benefit of ensuring
2984 that /dev only contains device nodes, directories and
2985 symlinks, and nothing else.
2987 * sd-daemon gained two new calls sd_pid_notify() and
2988 sd_pid_notifyf(). They are similar to sd_notify() and
2989 sd_notifyf(), but allow overriding of the source PID of
2990 notification messages if permissions permit this. This is
2991 useful to send notify messages on behalf of a different
2992 process (for example, the parent process). The
2993 systemd-notify tool has been updated to make use of this
2994 when sending messages (so that notification messages now
2995 originate from the shell script invoking systemd-notify and
2996 not the systemd-notify process itself. This should minimize
2997 a race where systemd fails to associate notification
2998 messages to services when the originating process already
3001 * A new "on-abnormal" setting for Restart= has been added. If
3002 set, it will result in automatic restarts on all "abnormal"
3003 reasons for a process to exit, which includes unclean
3004 signals, core dumps, timeouts and watchdog timeouts, but
3005 does not include clean and unclean exit codes or clean
3006 signals. Restart=on-abnormal is an alternative for
3007 Restart=on-failure for services that shall be able to
3008 terminate and avoid restarts on certain errors, by
3009 indicating so with an unclean exit code. Restart=on-failure
3010 or Restart=on-abnormal is now the recommended setting for
3011 all long-running services.
3013 * If the InaccessibleDirectories= service setting points to a
3014 mount point (or if there are any submounts contained within
3015 it), it is now attempted to completely unmount it, to make
3016 the file systems truly unavailable for the respective
3019 * The ReadOnlyDirectories= service setting and
3020 systemd-nspawn's --read-only parameter are now recursively
3021 applied to all submounts, too.
3023 * Mount units may now be created transiently via the bus APIs.
3025 * The support for SysV and LSB init scripts has been removed
3026 from the systemd daemon itself. Instead, it is now
3027 implemented as a generator that creates native systemd units
3028 from these scripts when needed. This enables us to remove a
3029 substantial amount of legacy code from PID 1, following the
3030 fact that many distributions only ship a very small number
3031 of LSB/SysV init scripts nowadays.
3033 * Privileged Xen (dom0) domains are not considered
3034 virtualization anymore by the virtualization detection
3035 logic. After all, they generally have unrestricted access to
3036 the hardware and usually are used to manage the unprivileged
3039 * systemd-tmpfiles gained a new "C" line type, for copying
3040 files or entire directories.
3042 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
3043 lines. So far, they have been non-globbing versions of the
3044 latter, and have thus been redundant. In future, it is
3045 recommended to only use "z". "m" has hence been removed
3046 from the documentation, even though it stays supported.
3048 * A tmpfiles snippet to recreate the most basic structure in
3049 /var has been added. This is enough to create the /var/run →
3050 /run symlink and create a couple of structural
3051 directories. This allows systems to boot up with an empty or
3052 volatile /var. Of course, while with this change, the core OS
3053 now is capable with dealing with a volatile /var, not all
3054 user services are ready for it. However, we hope that sooner
3055 or later, many service daemons will be changed upstream so
3056 that they are able to automatically create their necessary
3057 directories in /var at boot, should they be missing. This is
3058 the first step to allow state-less systems that only require
3059 the vendor image for /usr to boot.
3061 * systemd-nspawn has gained a new --tmpfs= switch to mount an
3062 empty tmpfs instance to a specific directory. This is
3063 particularly useful for making use of the automatic
3064 reconstruction of /var (see above), by passing --tmpfs=/var.
3066 * Access modes specified in tmpfiles snippets may now be
3067 prefixed with "~", which indicates that they shall be masked
3068 by whether the existing file or directory is currently
3069 writable, readable or executable at all. Also, if specified,
3070 the sgid/suid/sticky bits will be masked for all
3073 * A new passive target unit "network-pre.target" has been
3074 added which is useful for services that shall run before any
3075 network is configured, for example firewall scripts.
3077 * The "floppy" group that previously owned the /dev/fd*
3078 devices is no longer used. The "disk" group is now used
3079 instead. Distributions should probably deprecate usage of
3082 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
3083 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
3084 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
3085 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
3086 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
3087 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
3090 — Berlin, 2014-06-11
3094 * A new "systemd-timesyncd" daemon has been added for
3095 synchronizing the system clock across the network. It
3096 implements an SNTP client. In contrast to NTP
3097 implementations such as chrony or the NTP reference server,
3098 this only implements a client side, and does not bother with
3099 the full NTP complexity, focusing only on querying time from
3100 one remote server and synchronizing the local clock to
3101 it. Unless you intend to serve NTP to networked clients or
3102 want to connect to local hardware clocks, this simple NTP
3103 client should be more than appropriate for most
3104 installations. The daemon runs with minimal privileges, and
3105 has been hooked up with networkd to only operate when
3106 network connectivity is available. The daemon saves the
3107 current clock to disk every time a new NTP sync has been
3108 acquired, and uses this to possibly correct the system clock
3109 early at bootup, in order to accommodate for systems that
3110 lack an RTC such as the Raspberry Pi and embedded devices,
3111 and to make sure that time monotonically progresses on these
3112 systems, even if it is not always correct. To make use of
3113 this daemon, a new system user and group "systemd-timesync"
3114 needs to be created on installation of systemd.
3116 * The queue "seqnum" interface of libudev has been disabled, as
3117 it was generally incompatible with device namespacing as
3118 sequence numbers of devices go "missing" if the devices are
3119 part of a different namespace.
3121 * "systemctl list-timers" and "systemctl list-sockets" gained
3122 a --recursive switch for showing units of these types also
3123 for all local containers, similar in style to the already
3124 supported --recursive switch for "systemctl list-units".
3126 * A new RebootArgument= setting has been added for service
3127 units, which may be used to specify a kernel reboot argument
3128 to use when triggering reboots with StartLimitAction=.
3130 * A new FailureAction= setting has been added for service
3131 units which may be used to specify an operation to trigger
3132 when a service fails. This works similarly to
3133 StartLimitAction=, but unlike it, controls what is done
3134 immediately rather than only after several attempts to
3135 restart the service in question.
3137 * hostnamed got updated to also expose the kernel name,
3138 release, and version on the bus. This is useful for
3139 executing commands like hostnamectl with the -H switch.
3140 systemd-analyze makes use of this to properly display
3141 details when running non-locally.
3143 * The bootchart tool can now show cgroup information in the
3144 graphs it generates.
3146 * The CFS CPU quota cgroup attribute is now exposed for
3147 services. The new CPUQuota= switch has been added for this
3148 which takes a percentage value. Setting this will have the
3149 result that a service may never get more CPU time than the
3150 specified percentage, even if the machine is otherwise idle.
3152 * systemd-networkd learned IPIP and SIT tunnel support.
3154 * LSB init scripts exposing a dependency on $network will now
3155 get a dependency on network-online.target rather than simply
3156 network.target. This should bring LSB handling closer to
3157 what it was on SysV systems.
3159 * A new fsck.repair= kernel option has been added to control
3160 how fsck shall deal with unclean file systems at boot.
3162 * The (.ini) configuration file parser will now silently
3163 ignore sections whose name begins with "X-". This may be
3164 used to maintain application-specific extension sections in unit
3167 * machined gained a new API to query the IP addresses of
3168 registered containers. "machinectl status" has been updated
3169 to show these addresses in its output.
3171 * A new call sd_uid_get_display() has been added to the
3172 sd-login APIs for querying the "primary" session of a
3173 user. The "primary" session of the user is elected from the
3174 user's sessions and generally a graphical session is
3175 preferred over a text one.
3177 * A minimal systemd-resolved daemon has been added. It
3178 currently simply acts as a companion to systemd-networkd and
3179 manages resolv.conf based on per-interface DNS
3180 configuration, possibly supplied via DHCP. In the long run
3181 we hope to extend this into a local DNSSEC enabled DNS and
3184 * The systemd-networkd-wait-online tool is now enabled by
3185 default. It will delay network-online.target until a network
3186 connection has been configured. The tool primarily integrates
3187 with networkd, but will also make a best effort to make sense
3188 of network configuration performed in some other way.
3190 * Two new service options StartupCPUShares= and
3191 StartupBlockIOWeight= have been added that work similarly to
3192 CPUShares= and BlockIOWeight= however only apply during
3193 system startup. This is useful to prioritize certain services
3194 differently during bootup than during normal runtime.
3196 * hostnamed has been changed to prefer the statically
3197 configured hostname in /etc/hostname (unless set to
3198 'localhost' or empty) over any dynamic one supplied by
3199 dhcp. With this change, the rules for picking the hostname
3200 match more closely the rules of other configuration settings
3201 where the local administrator's configuration in /etc always
3202 overrides any other settings.
3204 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
3205 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
3206 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
3207 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
3208 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
3209 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
3210 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
3211 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
3212 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
3213 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
3214 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
3215 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
3216 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
3217 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
3218 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
3219 Lindskog, WaLyong Cho, Will Woods, Zbigniew
3222 — Beijing, 2014-05-28
3226 * When restoring the screen brightness at boot, stay away from
3227 the darkest setting or from the lowest 5% of the available
3228 range, depending on which is the larger value of both. This
3229 should effectively protect the user from rebooting into a
3230 black screen, should the brightness have been set to minimum
3233 * sd-login gained a new sd_machine_get_class() call to
3234 determine the class ("vm" or "container") of a machine
3235 registered with machined.
3237 * sd-login gained new calls
3238 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
3239 to query the identity of the peer of a local AF_UNIX
3240 connection. They operate similarly to their sd_pid_get_xyz()
3243 * PID 1 will now maintain a system-wide system state engine
3244 with the states "starting", "running", "degraded",
3245 "maintenance", "stopping". These states are bound to system
3246 startup, normal runtime, runtime with at least one failed
3247 service, rescue/emergency mode and system shutdown. This
3248 state is shown in the "systemctl status" output when no unit
3249 name is passed. It is useful to determine system state, in
3250 particularly when doing so for many systems or containers at
3253 * A new command "list-machines" has been added to "systemctl"
3254 that lists all local OS containers and shows their system
3255 state (see above), if systemd runs inside of them.
3257 * systemctl gained a new "-r" switch to recursively enumerate
3258 units on all local containers, when used with the
3259 "list-unit" command (which is the default one that is
3260 executed when no parameters are specified).
3262 * The GPT automatic partition discovery logic will now honour
3263 two GPT partition flags: one may be set on a partition to
3264 cause it to be mounted read-only, and the other may be set
3265 on a partition to ignore it during automatic discovery.
3267 * Two new GPT type UUIDs have been added for automatic root
3268 partition discovery, for 32-bit and 64-bit ARM. This is not
3269 particularly useful for discovering the root directory on
3270 these architectures during bare-metal boots (since UEFI is
3271 not common there), but still very useful to allow booting of
3272 ARM disk images in nspawn with the -i option.
3274 * MAC addresses of interfaces created with nspawn's
3275 --network-interface= switch will now be generated from the
3276 machine name, and thus be stable between multiple invocations
3279 * logind will now automatically remove all IPC objects owned
3280 by a user if she or he fully logs out. This makes sure that
3281 users who are logged out cannot continue to consume IPC
3282 resources. This covers SysV memory, semaphores and message
3283 queues as well as POSIX shared memory and message
3284 queues. Traditionally, SysV and POSIX IPC had no life-cycle
3285 limits. With this functionality, that is corrected. This may
3286 be turned off by using the RemoveIPC= switch of logind.conf.
3288 * The systemd-machine-id-setup and tmpfiles tools gained a
3289 --root= switch to operate on a specific root directory,
3292 * journald can now forward logged messages to the TTYs of all
3293 logged in users ("wall"). This is the default for all
3294 emergency messages now.
3296 * A new tool systemd-journal-remote has been added to stream
3297 journal log messages across the network.
3299 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
3300 controller trees are mounted into it. Note that the
3301 directories mounted beneath it are not read-only. This is a
3302 security measure and is particularly useful because glibc
3303 actually includes a search logic to pick any tmpfs it can
3304 find to implement shm_open() if /dev/shm is not available
3305 (which it might very well be in namespaced setups).
3307 * machinectl gained a new "poweroff" command to cleanly power
3308 down a local OS container.
3310 * The PrivateDevices= unit file setting will now also drop the
3311 CAP_MKNOD capability from the capability bound set, and
3312 imply DevicePolicy=closed.
3314 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
3315 comprehensively on all long-running systemd services where
3316 this is appropriate.
3318 * systemd-udevd will now run in a disassociated mount
3319 namespace. To mount directories from udev rules, make sure to
3320 pull in mount units via SYSTEMD_WANTS properties.
3322 * The kdbus support gained support for uploading policy into
3323 the kernel. sd-bus gained support for creating "monitoring"
3324 connections that can eavesdrop into all bus communication
3325 for debugging purposes.
3327 * Timestamps may now be specified in seconds since the UNIX
3328 epoch Jan 1st, 1970 by specifying "@" followed by the value
3331 * Native tcpwrap support in systemd has been removed. tcpwrap
3332 is old code, not really maintained anymore and has serious
3333 shortcomings, and better options such as firewalls
3334 exist. For setups that require tcpwrap usage, please
3335 consider invoking your socket-activated service via tcpd,
3336 like on traditional inetd.
3338 * A new system.conf configuration option
3339 DefaultTimerAccuracySec= has been added that controls the
3340 default AccuracySec= setting of .timer units.
3342 * Timer units gained a new WakeSystem= switch. If enabled,
3343 timers configured this way will cause the system to resume
3344 from system suspend (if the system supports that, which most
3347 * Timer units gained a new Persistent= switch. If enabled,
3348 timers configured this way will save to disk when they have
3349 been last triggered. This information is then used on next
3350 reboot to possible execute overdue timer events, that
3351 could not take place because the system was powered off.
3352 This enables simple anacron-like behaviour for timer units.
3354 * systemctl's "list-timers" will now also list the time a
3355 timer unit was last triggered in addition to the next time
3356 it will be triggered.
3358 * systemd-networkd will now assign predictable IPv4LL
3359 addresses to its local interfaces.
3361 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
3362 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
3363 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
3364 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
3365 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
3366 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
3367 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
3368 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
3371 — Berlin, 2014-03-25
3375 * A new unit file setting RestrictAddressFamilies= has been
3376 added to restrict which socket address families unit
3377 processes gain access to. This takes address family names
3378 like "AF_INET" or "AF_UNIX", and is useful to minimize the
3379 attack surface of services via exotic protocol stacks. This
3380 is built on seccomp system call filters.
3382 * Two new unit file settings RuntimeDirectory= and
3383 RuntimeDirectoryMode= have been added that may be used to
3384 manage a per-daemon runtime directories below /run. This is
3385 an alternative for setting up directory permissions with
3386 tmpfiles snippets, and has the advantage that the runtime
3387 directory's lifetime is bound to the daemon runtime and that
3388 the daemon starts up with an empty directory each time. This
3389 is particularly useful when writing services that drop
3390 privileges using the User= or Group= setting.
3392 * The DeviceAllow= unit setting now supports globbing for
3393 matching against device group names.
3395 * The systemd configuration file system.conf gained new
3396 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
3397 DefaultMemoryAccounting= to globally turn on/off accounting
3398 for specific resources (cgroups) for all units. These
3399 settings may still be overridden individually in each unit
3402 * systemd-gpt-auto-generator is now able to discover /srv and
3403 root partitions in addition to /home and swap partitions. It
3404 also supports LUKS-encrypted partitions now. With this in
3405 place, automatic discovery of partitions to mount following
3406 the Discoverable Partitions Specification
3407 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
3408 is now a lot more complete. This allows booting without
3409 /etc/fstab and without root= on the kernel command line on
3410 systems prepared appropriately.
3412 * systemd-nspawn gained a new --image= switch which allows
3413 booting up disk images and Linux installations on any block
3414 device that follow the Discoverable Partitions Specification
3415 (see above). This means that installations made with
3416 appropriately updated installers may now be started and
3417 deployed using container managers, completely
3418 unmodified. (We hope that libvirt-lxc will add support for
3419 this feature soon, too.)
3421 * systemd-nspawn gained a new --network-macvlan= setting to
3422 set up a private macvlan interface for the
3423 container. Similarly, systemd-networkd gained a new
3424 Kind=macvlan setting in .netdev files.
3426 * systemd-networkd now supports configuring local addresses
3429 * A new tool systemd-network-wait-online has been added to
3430 synchronously wait for network connectivity using
3433 * The sd-bus.h bus API gained a new sd_bus_track object for
3434 tracking the life-cycle of bus peers. Note that sd-bus.h is
3435 still not a public API though (unless you specify
3436 --enable-kdbus on the configure command line, which however
3437 voids your warranty and you get no API stability guarantee).
3439 * The $XDG_RUNTIME_DIR runtime directories for each user are
3440 now individual tmpfs instances, which has the benefit of
3441 introducing separate pools for each user, with individual
3442 size limits, and thus making sure that unprivileged clients
3443 can no longer negatively impact the system or other users by
3444 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
3445 RuntimeDirectorySize= has been introduced that allows
3446 controlling the default size limit for all users. It
3447 defaults to 10% of the available physical memory. This is no
3448 replacement for quotas on tmpfs though (which the kernel
3449 still does not support), as /dev/shm and /tmp are still
3450 shared resources used by both the system and unprivileged
3453 * logind will now automatically turn off automatic suspending
3454 on laptop lid close when more than one display is
3455 connected. This was previously expected to be implemented
3456 individually in desktop environments (such as GNOME),
3457 however has been added to logind now, in order to fix a
3458 boot-time race where a desktop environment might not have
3459 been started yet and thus not been able to take an inhibitor
3460 lock at the time where logind already suspends the system
3461 due to a closed lid.
3463 * logind will now wait at least 30s after each system
3464 suspend/resume cycle, and 3min after system boot before
3465 suspending the system due to a closed laptop lid. This
3466 should give USB docking stations and similar enough time to
3467 be probed and configured after system resume and boot in
3468 order to then act as suspend blocker.
3470 * systemd-run gained a new --property= setting which allows
3471 initialization of resource control properties (and others)
3472 for the created scope or service unit. Example: "systemd-run
3473 --property=BlockIOWeight=10 updatedb" may be used to run
3474 updatedb at a low block IO scheduling weight.
3476 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
3477 now also work in --scope mode.
3479 * When systemd is compiled with kdbus support, basic support
3480 for enforced policies is now in place. (Note that enabling
3481 kdbus still voids your warranty and no API compatibility
3484 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
3485 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
3486 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
3487 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
3488 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
3489 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
3490 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
3491 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
3492 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
3493 Zbigniew Jędrzejewski-Szmek
3495 — Berlin, 2014-03-12
3499 * systemd will now relabel /dev after loading the SMACK policy
3500 according to SMACK rules.
3502 * A new unit file option AppArmorProfile= has been added to
3503 set the AppArmor profile for the processes of a unit.
3505 * A new condition check ConditionArchitecture= has been added
3506 to conditionalize units based on the system architecture, as
3507 reported by uname()'s "machine" field.
3509 * systemd-networkd now supports matching on the system
3510 virtualization, architecture, kernel command line, host name
3513 * logind is now a lot more aggressive when suspending the
3514 machine due to a closed laptop lid. Instead of acting only
3515 on the lid close action, it will continuously watch the lid
3516 status and act on it. This is useful for laptops where the
3517 power button is on the outside of the chassis so that it can
3518 be reached without opening the lid (such as the Lenovo
3519 Yoga). On those machines, logind will now immediately
3520 re-suspend the machine if the power button has been
3521 accidentally pressed while the laptop was suspended and in a
3522 backpack or similar.
3524 * logind will now watch SW_DOCK switches and inhibit reaction
3525 to the lid switch if it is pressed. This means that logind
3526 will not suspend the machine anymore if the lid is closed
3527 and the system is docked, if the laptop supports SW_DOCK
3528 notifications via the input layer. Note that ACPI docking
3529 stations do not generate this currently. Also note that this
3530 logic is usually not fully sufficient and Desktop
3531 Environments should take a lid switch inhibitor lock when an
3532 external display is connected, as systemd will not watch
3535 * nspawn will now make use of the devices cgroup controller by
3536 default, and only permit creation of and access to the usual
3537 API device nodes like /dev/null or /dev/random, as well as
3538 access to (but not creation of) the pty devices.
3540 * We will now ship a default .network file for
3541 systemd-networkd that automatically configures DHCP for
3542 network interfaces created by nspawn's --network-veth or
3543 --network-bridge= switches.
3545 * systemd will now understand the usual M, K, G, T suffixes
3546 according to SI conventions (i.e. to the base 1000) when
3547 referring to throughput and hardware metrics. It will stay
3548 with IEC conventions (i.e. to the base 1024) for software
3549 metrics, according to what is customary according to
3550 Wikipedia. We explicitly document which base applies for
3551 each configuration option.
3553 * The DeviceAllow= setting in unit files now supports a syntax
3554 to whitelist an entire group of devices node majors at once,
3555 based on the /proc/devices listing. For example, with the
3556 string "char-pts", it is now possible to whitelist all
3557 current and future pseudo-TTYs at once.
3559 * sd-event learned a new "post" event source. Event sources of
3560 this type are triggered by the dispatching of any event
3561 source of a type that is not "post". This is useful for
3562 implementing clean-up and check event sources that are
3563 triggered by other work being done in the program.
3565 * systemd-networkd is no longer statically enabled, but uses
3566 the usual [Install] sections so that it can be
3567 enabled/disabled using systemctl. It still is enabled by
3570 * When creating a veth interface pair with systemd-nspawn, the
3571 host side will now be prefixed with "vb-" if
3572 --network-bridge= is used, and with "ve-" if --network-veth
3573 is used. This way, it is easy to distinguish these cases on
3574 the host, for example to apply different configuration to
3575 them with systemd-networkd.
3577 * The compatibility libraries for libsystemd-journal.so,
3578 libsystem-id128.so, libsystemd-login.so and
3579 libsystemd-daemon.so do not make use of IFUNC
3580 anymore. Instead, we now build libsystemd.so multiple times
3581 under these alternative names. This means that the footprint
3582 is drastically increased, but given that these are
3583 transitional compatibility libraries, this should not matter
3584 much. This change has been made necessary to support the ARM
3585 platform for these compatibility libraries, as the ARM
3586 toolchain is not really at the same level as the toolchain
3587 for other architectures like x86 and does not support
3588 IFUNC. Please make sure to use --enable-compat-libs only
3589 during a transitional period!
3591 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
3592 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
3593 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
3594 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
3595 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
3596 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
3597 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
3598 Zbigniew Jędrzejewski-Szmek
3600 — Berlin, 2014-02-24
3604 * A new component "systemd-networkd" has been added that can
3605 be used to configure local network interfaces statically or
3606 via DHCP. It is capable of bringing up bridges, VLANs, and
3607 bonding. Currently, no hook-ups for interactive network
3608 configuration are provided. Use this for your initrd,
3609 container, embedded, or server setup if you need a simple,
3610 yet powerful, network configuration solution. This
3611 configuration subsystem is quite nifty, as it allows wildcard
3612 hotplug matching in interfaces. For example, with a single
3613 configuration snippet, you can configure that all Ethernet
3614 interfaces showing up are automatically added to a bridge,
3615 or similar. It supports link-sensing and more.
3617 * A new tool "systemd-socket-proxyd" has been added which can
3618 act as a bidirectional proxy for TCP sockets. This is
3619 useful for adding socket activation support to services that
3620 do not actually support socket activation, including virtual
3621 machines and the like.
3623 * Add a new tool to save/restore rfkill state on
3626 * Save/restore state of keyboard backlights in addition to
3627 display backlights on shutdown/boot.
3629 * udev learned a new SECLABEL{} construct to label device
3630 nodes with a specific security label when they appear. For
3631 now, only SECLABEL{selinux} is supported, but the syntax is
3632 prepared for additional security frameworks.
3634 * udev gained a new scheme to configure link-level attributes
3635 from files in /etc/systemd/network/*.link. These files can
3636 match against MAC address, device path, driver name and type,
3637 and will apply attributes like the naming policy, link speed,
3638 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
3639 address assignment policy (randomized, ...).
3641 * The configuration of network interface naming rules for
3642 "permanent interface names" has changed: a new NamePolicy=
3643 setting in the [Link] section of .link files determines the
3644 priority of possible naming schemes (onboard, slot, MAC,
3645 path). The default value of this setting is determined by
3646 /usr/lib/net/links/99-default.link. Old
3647 80-net-name-slot.rules udev configuration file has been
3648 removed, so local configuration overriding this file should
3649 be adapated to override 99-default.link instead.
3651 * When the User= switch is used in a unit file, also
3652 initialize $SHELL= based on the user database entry.
3654 * systemd no longer depends on libdbus. All communication is
3655 now done with sd-bus, systemd's low-level bus library
3658 * kdbus support has been added to PID 1 itself. When kdbus is
3659 enabled, this causes PID 1 to set up the system bus and
3660 enable support for a new ".busname" unit type that
3661 encapsulates bus name activation on kdbus. It works a little
3662 bit like ".socket" units, except for bus names. A new
3663 generator has been added that converts classic dbus1 service
3664 activation files automatically into native systemd .busname
3667 * sd-bus: add a light-weight vtable implementation that allows
3668 defining objects on the bus with a simple static const
3669 vtable array of its methods, signals and properties.
3671 * systemd will not generate or install static dbus
3672 introspection data anymore to /usr/share/dbus-1/interfaces,
3673 as the precise format of these files is unclear, and
3674 nothing makes use of it.
3676 * A proxy daemon is now provided to proxy clients connecting
3677 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
3678 compatibility with classic D-Bus.
3680 * A bus driver implementation has been added that supports the
3681 classic D-Bus bus driver calls on kdbus, also for
3682 compatibility purposes.
3684 * A new API "sd-event.h" has been added that implements a
3685 minimal event loop API built around epoll. It provides a
3686 couple of features that direct epoll usage is lacking:
3687 prioritization of events, scales to large numbers of timer
3688 events, per-event timer slack (accuracy), system-wide
3689 coalescing of timer events, exit handlers, watchdog
3690 supervision support using systemd's sd_notify() API, child
3693 * A new API "sd-rntl.h" has been added that provides an API
3694 around the route netlink interface of the kernel, similar in
3695 style to "sd-bus.h".
3697 * A new API "sd-dhcp-client.h" has been added that provides a
3698 small DHCPv4 client-side implementation. This is used by
3701 * There is a new kernel command line option
3702 "systemd.restore_state=0|1". When set to "0", none of the
3703 systemd tools will restore saved runtime state to hardware
3704 devices. More specifically, the rfkill and backlight states
3707 * The FsckPassNo= compatibility option in mount/service units
3708 has been removed. The fstab generator will now add the
3709 necessary dependencies automatically, and does not require
3710 PID1's support for that anymore.
3712 * journalctl gained a new switch, --list-boots, that lists
3713 recent boots with their times and boot IDs.
3715 * The various tools like systemctl, loginctl, timedatectl,
3716 busctl, systemd-run, ... have gained a new switch "-M" to
3717 connect to a specific, local OS container (as direct
3718 connection, without requiring SSH). This works on any
3719 container that is registered with machined, such as those
3720 created by libvirt-lxc or nspawn.
3722 * systemd-run and systemd-analyze also gained support for "-H"
3723 to connect to remote hosts via SSH. This is particularly
3724 useful for systemd-run because it enables queuing of jobs
3725 onto remote systems.
3727 * machinectl gained a new command "login" to open a getty
3728 login in any local container. This works with any container
3729 that is registered with machined (such as those created by
3730 libvirt-lxc or nspawn), and which runs systemd inside.
3732 * machinectl gained a new "reboot" command that may be used to
3733 trigger a reboot on a specific container that is registered
3734 with machined. This works on any container that runs an init
3735 system of some kind.
3737 * systemctl gained a new "list-timers" command to print a nice
3738 listing of installed timer units with the times they elapse
3741 * Alternative reboot() parameters may now be specified on the
3742 "systemctl reboot" command line and are passed to the
3743 reboot() system call.
3745 * systemctl gained a new --job-mode= switch to configure the
3746 mode to queue a job with. This is a more generic version of
3747 --fail, --irreversible, and --ignore-dependencies, which are
3748 still available but not advertised anymore.
3750 * /etc/systemd/system.conf gained new settings to configure
3751 various default timeouts of units, as well as the default
3752 start limit interval and burst. These may still be overridden
3755 * PID1 will now export on the bus profile data of the security
3756 policy upload process (such as the SELinux policy upload to
3759 * journald: when forwarding logs to the console, include
3760 timestamps (following the setting in
3761 /sys/module/printk/parameters/time).
3763 * OnCalendar= in timer units now understands the special
3764 strings "yearly" and "annually". (Both are equivalent)
3766 * The accuracy of timer units is now configurable with the new
3767 AccuracySec= setting. It defaults to 1min.
3769 * A new dependency type JoinsNamespaceOf= has been added that
3770 allows running two services within the same /tmp and network
3771 namespace, if PrivateNetwork= or PrivateTmp= are used.
3773 * A new command "cat" has been added to systemctl. It outputs
3774 the original unit file of a unit, and concatenates the
3775 contents of additional "drop-in" unit file snippets, so that
3776 the full configuration is shown.
3778 * systemctl now supports globbing on the various "list-xyz"
3779 commands, like "list-units" or "list-sockets", as well as on
3780 those commands which take multiple unit names.
3782 * journalctl's --unit= switch gained support for globbing.
3784 * All systemd daemons now make use of the watchdog logic so
3785 that systemd automatically notices when they hang.
3787 * If the $container_ttys environment variable is set,
3788 getty-generator will automatically spawn a getty for each
3789 listed tty. This is useful for container managers to request
3790 login gettys to be spawned on as many ttys as needed.
3792 * %h, %s, %U specifier support is not available anymore when
3793 used in unit files for PID 1. This is because NSS calls are
3794 not safe from PID 1. They stay available for --user
3795 instances of systemd, and as special case for the root user.
3797 * loginctl gained a new "--no-legend" switch to turn off output
3800 * The "sd-login.h" API gained three new calls:
3801 sd_session_is_remote(), sd_session_get_remote_user(),
3802 sd_session_get_remote_host() to query information about
3805 * The udev hardware database now also carries vendor/product
3806 information of SDIO devices.
3808 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
3809 determine whether watchdog notifications are requested by
3812 * Socket-activated per-connection services now include a
3813 short description of the connection parameters in the
3816 * tmpfiles gained a new "--boot" option. When this is not used,
3817 only lines where the command character is not suffixed with
3818 "!" are executed. When this option is specified, those
3819 options are executed too. This partitions tmpfiles
3820 directives into those that can be safely executed at any
3821 time, and those which should be run only at boot (for
3822 example, a line that creates /run/nologin).
3824 * A new API "sd-resolve.h" has been added which provides a simple
3825 asynchronous wrapper around glibc NSS host name resolution
3826 calls, such as getaddrinfo(). In contrast to glibc's
3827 getaddrinfo_a(), it does not use signals. In contrast to most
3828 other asynchronous name resolution libraries, this one does
3829 not reimplement DNS, but reuses NSS, so that alternate
3830 host name resolution systems continue to work, such as mDNS,
3831 LDAP, etc. This API is based on libasyncns, but it has been
3832 cleaned up for inclusion in systemd.
3834 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
3835 "sd-daemon.h" are no longer found in individual libraries
3836 libsystemd-journal.so, libsystemd-login.so,
3837 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
3838 merged them into a single library, libsystemd.so, which
3839 provides all symbols. The reason for this is cyclic
3840 dependencies, as these libraries tend to use each other's
3841 symbols. So far, we have managed to workaround that by linking
3842 a copy of a good part of our code into each of these
3843 libraries again and again, which, however, makes certain
3844 things hard to do, like sharing static variables. Also, it
3845 substantially increases footprint. With this change, there
3846 is only one library for the basic APIs systemd
3847 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
3848 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
3849 library as well, however are subject to the --enable-kdbus
3850 switch (see below). Note that "sd-dhcp-client.h" is not part
3851 of this library (this is because it only consumes, never
3852 provides, services of/to other APIs). To make the transition
3853 easy from the separate libraries to the unified one, we
3854 provide the --enable-compat-libs compile-time switch which
3855 will generate stub libraries that are compatible with the
3856 old ones but redirect all calls to the new one.
3858 * All of the kdbus logic and the new APIs "sd-bus.h",
3859 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
3860 and "sd-utf8.h" are compile-time optional via the
3861 "--enable-kdbus" switch, and they are not compiled in by
3862 default. To make use of kdbus, you have to explicitly enable
3863 the switch. Note however, that neither the kernel nor the
3864 userspace API for all of this is considered stable yet. We
3865 want to maintain the freedom to still change the APIs for
3866 now. By specifying this build-time switch, you acknowledge
3867 that you are aware of the instability of the current
3870 * Also, note that while kdbus is pretty much complete,
3871 it lacks one thing: proper policy support. This means you
3872 can build a fully working system with all features; however,
3873 it will be highly insecure. Policy support will be added in
3874 one of the next releases, at the same time that we will
3875 declare the APIs stable.
3877 * When the kernel command line argument "kdbus" is specified,
3878 systemd will automatically load the kdbus.ko kernel module. At
3879 this stage of development, it is only useful for testing kdbus
3880 and should not be used in production. Note: if "--enable-kdbus"
3881 is specified, and the kdbus.ko kernel module is available, and
3882 "kdbus" is added to the kernel command line, the entire system
3883 runs with kdbus instead of dbus-daemon, with the above mentioned
3884 problem of missing the system policy enforcement. Also a future
3885 version of kdbus.ko or a newer systemd will not be compatible with
3886 each other, and will unlikely be able to boot the machine if only
3887 one of them is updated.
3889 * systemctl gained a new "import-environment" command which
3890 uploads the caller's environment (or parts thereof) into the
3891 service manager so that it is inherited by services started
3892 by the manager. This is useful to upload variables like
3893 $DISPLAY into the user service manager.
3895 * A new PrivateDevices= switch has been added to service units
3896 which allows running a service with a namespaced /dev
3897 directory that does not contain any device nodes for
3898 physical devices. More specifically, it only includes devices
3899 such as /dev/null, /dev/urandom, and /dev/zero which are API
3902 * logind has been extended to support behaviour like VT
3903 switching on seats that do not support a VT. This makes
3904 multi-session available on seats that are not the first seat
3905 (seat0), and on systems where kernel support for VTs has
3906 been disabled at compile-time.
3908 * If a process holds a delay lock for system sleep or shutdown
3909 and fails to release it in time, we will now log its
3910 identity. This makes it easier to identify processes that
3911 cause slow suspends or power-offs.
3913 * When parsing /etc/crypttab, support for a new key-slot=
3914 option as supported by Debian is added. It allows indicating
3915 which LUKS slot to use on disk, speeding up key loading.
3917 * The sd_journald_sendv() API call has been checked and
3918 officially declared to be async-signal-safe so that it may
3919 be invoked from signal handlers for logging purposes.
3921 * Boot-time status output is now enabled automatically after a
3922 short timeout if boot does not progress, in order to give
3923 the user an indication what she or he is waiting for.
3925 * The boot-time output has been improved to show how much time
3926 remains until jobs expire.
3928 * The KillMode= switch in service units gained a new possible
3929 value "mixed". If set, and the unit is shut down, then the
3930 initial SIGTERM signal is sent only to the main daemon
3931 process, while the following SIGKILL signal is sent to
3932 all remaining processes of the service.
3934 * When a scope unit is registered, a new property "Controller"
3935 may be set. If set to a valid bus name, systemd will send a
3936 RequestStop() signal to this name when it would like to shut
3937 down the scope. This may be used to hook manager logic into
3938 the shutdown logic of scope units. Also, scope units may now
3939 be put in a special "abandoned" state, in which case the
3940 manager process which created them takes no further
3941 responsibilities for it.
3943 * When reading unit files, systemd will now verify
3944 the access mode of these files, and warn about certain
3945 suspicious combinations. This has been added to make it
3946 easier to track down packaging bugs where unit files are
3947 marked executable or world-writable.
3949 * systemd-nspawn gained a new "--setenv=" switch to set
3950 container-wide environment variables. The similar option in
3951 systemd-activate was renamed from "--environment=" to
3952 "--setenv=" for consistency.
3954 * systemd-nspawn has been updated to create a new kdbus domain
3955 for each container that is invoked, thus allowing each
3956 container to have its own set of system and user buses,
3957 independent of the host.
3959 * systemd-nspawn gained a new --drop-capability= switch to run
3960 the container with less capabilities than the default. Both
3961 --drop-capability= and --capability= now take the special
3962 string "all" for dropping or keeping all capabilities.
3964 * systemd-nspawn gained new switches for executing containers
3965 with specific SELinux labels set.
3967 * systemd-nspawn gained a new --quiet switch to not generate
3968 any additional output but the container's own console
3971 * systemd-nspawn gained a new --share-system switch to run a
3972 container without PID namespacing enabled.
3974 * systemd-nspawn gained a new --register= switch to control
3975 whether the container is registered with systemd-machined or
3976 not. This is useful for containers that do not run full
3977 OS images, but only specific apps.
3979 * systemd-nspawn gained a new --keep-unit which may be used
3980 when invoked as the only program from a service unit, and
3981 results in registration of the unit service itself in
3982 systemd-machined, instead of a newly opened scope unit.
3984 * systemd-nspawn gained a new --network-interface= switch for
3985 moving arbitrary interfaces to the container. The new
3986 --network-veth switch creates a virtual Ethernet connection
3987 between host and container. The new --network-bridge=
3988 switch then allows assigning the host side of this virtual
3989 Ethernet connection to a bridge device.
3991 * systemd-nspawn gained a new --personality= switch for
3992 setting the kernel personality for the container. This is
3993 useful when running a 32-bit container on a 64-bit host. A
3994 similar option Personality= is now also available for service
3997 * logind will now also track a "Desktop" identifier for each
3998 session which encodes the desktop environment of it. This is
3999 useful for desktop environments that want to identify
4000 multiple running sessions of itself easily.
4002 * A new SELinuxContext= setting for service units has been
4003 added that allows setting a specific SELinux execution
4004 context for a service.
4006 * Most systemd client tools will now honour $SYSTEMD_LESS for
4007 settings of the "less" pager. By default, these tools will
4008 override $LESS to allow certain operations to work, such as
4009 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
4010 influence this logic.
4012 * systemd's "seccomp" hook-up has been changed to make use of
4013 the libseccomp library instead of using its own
4014 implementation. This has benefits for portability among
4017 * For usage together with SystemCallFilter=, a new
4018 SystemCallErrorNumber= setting has been introduced that
4019 allows configuration of a system error number to be returned
4020 on filtered system calls, instead of immediately killing the
4021 process. Also, SystemCallArchitectures= has been added to
4022 limit access to system calls of a particular architecture
4023 (in order to turn off support for unused secondary
4024 architectures). There is also a global
4025 SystemCallArchitectures= setting in system.conf now to turn
4026 off support for non-native system calls system-wide.
4028 * systemd requires a kernel with a working name_to_handle_at(),
4029 please see the kernel config requirements in the README file.
4031 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
4032 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
4033 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
4034 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
4035 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
4036 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
4037 Elia Pinto, Florian Weimer, George McCollister, Goffredo
4038 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
4039 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
4040 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
4041 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
4042 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
4043 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
4044 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
4045 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
4046 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
4047 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
4048 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
4049 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
4050 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
4051 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
4052 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
4053 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
4054 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
4056 — Berlin, 2014-02-20
4060 * logind has gained support for facilitating privileged input
4061 and drm device access for unprivileged clients. This work is
4062 useful to allow Wayland display servers (and similar
4063 programs, such as kmscon) to run under the user's ID and
4064 access input and drm devices which are normally
4065 protected. When this is used (and the kernel is new enough)
4066 logind will "mute" IO on the file descriptors passed to
4067 Wayland as long as it is in the background and "unmute" it
4068 if it returns into the foreground. This allows secure
4069 session switching without allowing background sessions to
4070 eavesdrop on input and display data. This also introduces
4071 session switching support if VT support is turned off in the
4072 kernel, and on seats that are not seat0.
4074 * A new kernel command line option luks.options= is understood
4075 now which allows specifying LUKS options for usage for LUKS
4076 encrypted partitions specified with luks.uuid=.
4078 * tmpfiles.d(5) snippets may now use specifier expansion in
4079 path names. More specifically %m, %b, %H, %v, are now
4080 replaced by the local machine id, boot id, hostname, and
4081 kernel version number.
4083 * A new tmpfiles.d(5) command "m" has been introduced which
4084 may be used to change the owner/group/access mode of a file
4085 or directory if it exists, but do nothing if it does not.
4087 * This release removes high-level support for the
4088 MemorySoftLimit= cgroup setting. The underlying kernel
4089 cgroup attribute memory.soft_limit= is currently badly
4090 designed and likely to be removed from the kernel API in its
4091 current form, hence we should not expose it for now.
4093 * The memory.use_hierarchy cgroup attribute is now enabled for
4094 all cgroups systemd creates in the memory cgroup
4095 hierarchy. This option is likely to be come the built-in
4096 default in the kernel anyway, and the non-hierarchical mode
4097 never made much sense in the intrinsically hierarchical
4100 * A new field _SYSTEMD_SLICE= is logged along with all journal
4101 messages containing the slice a message was generated
4102 from. This is useful to allow easy per-customer filtering of
4103 logs among other things.
4105 * systemd-journald will no longer adjust the group of journal
4106 files it creates to the "systemd-journal" group. Instead we
4107 rely on the journal directory to be owned by the
4108 "systemd-journal" group, and its setgid bit set, so that the
4109 kernel file system layer will automatically enforce that
4110 journal files inherit this group assignment. The reason for
4111 this change is that we cannot allow NSS look-ups from
4112 journald which would be necessary to resolve
4113 "systemd-journal" to a numeric GID, because this might
4114 create deadlocks if NSS involves synchronous queries to
4115 other daemons (such as nscd, or sssd) which in turn are
4116 logging clients of journald and might block on it, which
4117 would then dead lock. A tmpfiles.d(5) snippet included in
4118 systemd will make sure the setgid bit and group are
4119 properly set on the journal directory if it exists on every
4120 boot. However, we recommend adjusting it manually after
4121 upgrades too (or from RPM scriptlets), so that the change is
4122 not delayed until next reboot.
4124 * Backlight and random seed files in /var/lib/ have moved into
4125 the /var/lib/systemd/ directory, in order to centralize all
4126 systemd generated files in one directory.
4128 * Boot time performance measurements (as displayed by
4129 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
4130 performance information if that's available to determine how
4131 much time BIOS and boot loader initialization required. With
4132 a sufficiently new BIOS you hence no longer need to boot
4133 with Gummiboot to get access to such information.
4135 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
4136 Cristian Rodríguez, Dave Reisner, David Herrmann, David
4137 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
4138 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
4139 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
4140 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
4141 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
4143 — Berlin, 2013-10-02
4147 * The Restart= option for services now understands a new
4148 on-watchdog setting, which will restart the service
4149 automatically if the service stops sending out watchdog keep
4150 alive messages (as configured with WatchdogSec=).
4152 * The getty generator (which is responsible for bringing up a
4153 getty on configured serial consoles) will no longer only
4154 start a getty on the primary kernel console but on all
4155 others, too. This makes the order in which console= is
4156 specified on the kernel command line less important.
4158 * libsystemd-logind gained a new sd_session_get_vt() call to
4159 retrieve the VT number of a session.
4161 * If the option "tries=0" is set for an entry of /etc/crypttab
4162 its passphrase is queried indefinitely instead of any
4163 maximum number of tries.
4165 * If a service with a configure PID file terminates its PID
4166 file will now be removed automatically if it still exists
4167 afterwards. This should put an end to stale PID files.
4169 * systemd-run will now also take relative binary path names
4170 for execution and no longer insists on absolute paths.
4172 * InaccessibleDirectories= and ReadOnlyDirectories= now take
4173 paths that are optionally prefixed with "-" to indicate that
4174 it should not be considered a failure if they do not exist.
4176 * journalctl -o (and similar commands) now understands a new
4177 output mode "short-precise", it is similar to "short" but
4178 shows timestamps with usec accuracy.
4180 * The option "discard" (as known from Debian) is now
4181 synonymous to "allow-discards" in /etc/crypttab. In fact,
4182 "discard" is preferred now (since it is easier to remember
4185 * Some licensing clean-ups were made, so that more code is now
4186 LGPL-2.1 licensed than before.
4188 * A minimal tool to save/restore the display backlight
4189 brightness across reboots has been added. It will store the
4190 backlight setting as late as possible at shutdown, and
4191 restore it as early as possible during reboot.
4193 * A logic to automatically discover and enable home and swap
4194 partitions on GPT disks has been added. With this in place
4195 /etc/fstab becomes optional for many setups as systemd can
4196 discover certain partitions located on the root disk
4197 automatically. Home partitions are recognized under their
4198 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
4199 partitions are recognized under their GPT type ID
4200 0657fd6da4ab43c484e50933c84b4f4f.
4202 * systemd will no longer pass any environment from the kernel
4203 or initrd to system services. If you want to set an
4204 environment for all services, do so via the kernel command
4205 line systemd.setenv= assignment.
4207 * The systemd-sysctl tool no longer natively reads the file
4208 /etc/sysctl.conf. If desired, the file should be symlinked
4209 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
4210 legacy support by a symlink rather than built-in code, it
4211 also makes the otherwise hidden order of application of the
4212 different files visible. (Note that this partly reverts to a
4213 pre-198 application order of sysctl knobs!)
4215 * The "systemctl set-log-level" and "systemctl dump" commands
4216 have been moved to systemd-analyze.
4218 * systemd-run learned the new --remain-after-exit switch,
4219 which causes the scope unit not to be cleaned up
4220 automatically after the process terminated.
4222 * tmpfiles learned a new --exclude-prefix= switch to exclude
4223 certain paths from operation.
4225 * journald will now automatically flush all messages to disk
4226 as soon as a message at the log level CRIT, ALERT or EMERG
4229 Contributions from: Andrew Cook, Brandon Philips, Christian
4230 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
4231 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
4232 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
4233 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
4234 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
4235 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
4236 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
4237 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
4238 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
4239 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
4240 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
4241 William Giokas, Zbigniew Jędrzejewski-Szmek
4243 — Berlin, 2013-09-13
4247 * The documentation has been updated to cover the various new
4248 concepts introduced with 205.
4250 * Unit files now understand the new %v specifier which
4251 resolves to the kernel version string as returned by "uname
4254 * systemctl now supports filtering the unit list output by
4255 load state, active state and sub state, using the new
4258 * "systemctl status" will now show the results of the
4259 condition checks (like ConditionPathExists= and similar) of
4260 the last start attempts of the unit. They are also logged to
4263 * "journalctl -b" may now be used to look for boot output of a
4264 specific boot. Try "journalctl -b -1" for the previous boot,
4265 but the syntax is substantially more powerful.
4267 * "journalctl --show-cursor" has been added which prints the
4268 cursor string the last shown log line. This may then be used
4269 with the new "journalctl --after-cursor=" switch to continue
4270 browsing logs from that point on.
4272 * "journalctl --force" may now be used to force regeneration
4275 * Creation of "dead" device nodes has been moved from udev
4276 into kmod and tmpfiles. Previously, udev would read the kmod
4277 databases to pre-generate dead device nodes based on meta
4278 information contained in kernel modules, so that these would
4279 be auto-loaded on access rather then at boot. As this
4280 does not really have much to do with the exposing actual
4281 kernel devices to userspace this has always been slightly
4282 alien in the udev codebase. Following the new scheme kmod
4283 will now generate a runtime snippet for tmpfiles from the
4284 module meta information and it now is tmpfiles' job to the
4285 create the nodes. This also allows overriding access and
4286 other parameters for the nodes using the usual tmpfiles
4287 facilities. As side effect this allows us to remove the
4288 CAP_SYS_MKNOD capability bit from udevd entirely.
4290 * logind's device ACLs may now be applied to these "dead"
4291 devices nodes too, thus finally allowing managed access to
4292 devices such as /dev/snd/sequencer whithout loading the
4293 backing module right-away.
4295 * A new RPM macro has been added that may be used to apply
4296 tmpfiles configuration during package installation.
4298 * systemd-detect-virt and ConditionVirtualization= now can
4299 detect User-Mode-Linux machines (UML).
4301 * journald will now implicitly log the effective capabilities
4302 set of processes in the message metadata.
4304 * systemd-cryptsetup has gained support for TrueCrypt volumes.
4306 * The initrd interface has been simplified (more specifically,
4307 support for passing performance data via environment
4308 variables and fsck results via files in /run has been
4309 removed). These features were non-essential, and are
4310 nowadays available in a much nicer way by having systemd in
4311 the initrd serialize its state and have the hosts systemd
4312 deserialize it again.
4314 * The udev "keymap" data files and tools to apply keyboard
4315 specific mappings of scan to key codes, and force-release
4316 scan code lists have been entirely replaced by a udev
4317 "keyboard" builtin and a hwdb data file.
4319 * systemd will now honour the kernel's "quiet" command line
4320 argument also during late shutdown, resulting in a
4321 completely silent shutdown when used.
4323 * There's now an option to control the SO_REUSEPORT socket
4324 option in .socket units.
4326 * Instance units will now automatically get a per-template
4327 subslice of system.slice unless something else is explicitly
4328 configured. For example, instances of sshd@.service will now
4329 implicitly be placed in system-sshd.slice rather than
4330 system.slice as before.
4332 * Test coverage support may now be enabled at build time.
4334 Contributions from: Dave Reisner, Frederic Crozat, Harald
4335 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
4336 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
4337 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
4338 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
4339 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
4340 Giokas, Zbigniew Jędrzejewski-Szmek
4342 — Berlin, 2013-07-23
4346 * Two new unit types have been introduced:
4348 Scope units are very similar to service units, however, are
4349 created out of pre-existing processes — instead of PID 1
4350 forking off the processes. By using scope units it is
4351 possible for system services and applications to group their
4352 own child processes (worker processes) in a powerful way
4353 which then maybe used to organize them, or kill them
4354 together, or apply resource limits on them.
4356 Slice units may be used to partition system resources in an
4357 hierarchical fashion and then assign other units to them. By
4358 default there are now three slices: system.slice (for all
4359 system services), user.slice (for all user sessions),
4360 machine.slice (for VMs and containers).
4362 Slices and scopes have been introduced primarily in
4363 context of the work to move cgroup handling to a
4364 single-writer scheme, where only PID 1
4365 creates/removes/manages cgroups.
4367 * There's a new concept of "transient" units. In contrast to
4368 normal units these units are created via an API at runtime,
4369 not from configuration from disk. More specifically this
4370 means it is now possible to run arbitrary programs as
4371 independent services, with all execution parameters passed
4372 in via bus APIs rather than read from disk. Transient units
4373 make systemd substantially more dynamic then it ever was,
4374 and useful as a general batch manager.
4376 * logind has been updated to make use of scope and slice units
4377 for managing user sessions. As a user logs in he will get
4378 his own private slice unit, to which all sessions are added
4379 as scope units. We also added support for automatically
4380 adding an instance of user@.service for the user into the
4381 slice. Effectively logind will no longer create cgroup
4382 hierarchies on its own now, it will defer entirely to PID 1
4383 for this by means of scope, service and slice units. Since
4384 user sessions this way become entities managed by PID 1
4385 the output of "systemctl" is now a lot more comprehensive.
4387 * A new mini-daemon "systemd-machined" has been added which
4388 may be used by virtualization managers to register local
4389 VMs/containers. nspawn has been updated accordingly, and
4390 libvirt will be updated shortly. machined will collect a bit
4391 of meta information about the VMs/containers, and assign
4392 them their own scope unit (see above). The collected
4393 meta-data is then made available via the "machinectl" tool,
4394 and exposed in "ps" and similar tools. machined/machinectl
4395 is compile-time optional.
4397 * As discussed earlier, the low-level cgroup configuration
4398 options ControlGroup=, ControlGroupModify=,
4399 ControlGroupPersistent=, ControlGroupAttribute= have been
4400 removed. Please use high-level attribute settings instead as
4401 well as slice units.
4403 * A new bus call SetUnitProperties() has been added to alter
4404 various runtime parameters of a unit. This is primarily
4405 useful to alter cgroup parameters dynamically in a nice way,
4406 but will be extended later on to make more properties
4407 modifiable at runtime. systemctl gained a new set-properties
4408 command that wraps this call.
4410 * A new tool "systemd-run" has been added which can be used to
4411 run arbitrary command lines as transient services or scopes,
4412 while configuring a number of settings via the command
4413 line. This tool is currently very basic, however already
4414 very useful. We plan to extend this tool to even allow
4415 queuing of execution jobs with time triggers from the
4416 command line, similar in fashion to "at".
4418 * nspawn will now inform the user explicitly that kernels with
4419 audit enabled break containers, and suggest the user to turn
4422 * Support for detecting the IMA and AppArmor security
4423 frameworks with ConditionSecurity= has been added.
4425 * journalctl gained a new "-k" switch for showing only kernel
4426 messages, mimicking dmesg output; in addition to "--user"
4427 and "--system" switches for showing only user's own logs
4430 * systemd-delta can now show information about drop-in
4431 snippets extending unit files.
4433 * libsystemd-bus has been substantially updated but is still
4434 not available as public API.
4436 * systemd will now look for the "debug" argument on the kernel
4437 command line and enable debug logging, similar to what
4438 "systemd.log_level=debug" already did before.
4440 * "systemctl set-default", "systemctl get-default" has been
4441 added to configure the default.target symlink, which
4442 controls what to boot into by default.
4444 * "systemctl set-log-level" has been added as a convenient
4445 way to raise and lower systemd logging threshold.
4447 * "systemd-analyze plot" will now show the time the various
4448 generators needed for execution, as well as information
4449 about the unit file loading.
4451 * libsystemd-journal gained a new sd_journal_open_files() call
4452 for opening specific journal files. journactl also gained a
4453 new switch to expose this new functionality. Previously we
4454 only supported opening all files from a directory, or all
4455 files from the system, as opening individual files only is
4456 racy due to journal file rotation.
4458 * systemd gained the new DefaultEnvironment= setting in
4459 /etc/systemd/system.conf to set environment variables for
4462 * If a privileged process logs a journal message with the
4463 OBJECT_PID= field set, then journald will automatically
4464 augment this with additional OBJECT_UID=, OBJECT_GID=,
4465 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
4466 system services want to log events about specific client
4467 processes. journactl/systemctl has been updated to make use
4468 of this information if all log messages regarding a specific
4471 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
4472 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
4473 Reisner, David Coppa, David King, David Strauss, Eelco
4474 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
4475 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
4476 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
4477 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
4478 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
4479 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
4480 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
4481 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
4482 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
4483 Łukasz Stelmach, 장동준
4487 * The Python bindings gained some minimal support for the APIs
4488 exposed by libsystemd-logind.
4490 * ConditionSecurity= gained support for detecting SMACK. Since
4491 this condition already supports SELinux and AppArmor we only
4492 miss IMA for this. Patches welcome!
4494 Contributions from: Karol Lewandowski, Lennart Poettering,
4495 Zbigniew Jędrzejewski-Szmek
4499 * systemd-nspawn will now create /etc/resolv.conf if
4500 necessary, before bind-mounting the host's file onto it.
4502 * systemd-nspawn will now store meta information about a
4503 container on the container's cgroup as extended attribute
4504 fields, including the root directory.
4506 * The cgroup hierarchy has been reworked in many ways. All
4507 objects any of the components systemd creates in the cgroup
4508 tree are now suffixed. More specifically, user sessions are
4509 now placed in cgroups suffixed with ".session", users in
4510 cgroups suffixed with ".user", and nspawn containers in
4511 cgroups suffixed with ".nspawn". Furthermore, all cgroup
4512 names are now escaped in a simple scheme to avoid collision
4513 of userspace object names with kernel filenames. This work
4514 is preparation for making these objects relocatable in the
4515 cgroup tree, in order to allow easy resource partitioning of
4516 these objects without causing naming conflicts.
4518 * systemctl list-dependencies gained the new switches
4519 --plain, --reverse, --after and --before.
4521 * systemd-inhibit now shows the process name of processes that
4522 have taken an inhibitor lock.
4524 * nss-myhostname will now also resolve "localhost"
4525 implicitly. This makes /etc/hosts an optional file and
4526 nicely handles that on IPv6 ::1 maps to both "localhost" and
4529 * libsystemd-logind.so gained a new call
4530 sd_get_machine_names() to enumerate running containers and
4531 VMs (currently only supported by very new libvirt and
4532 nspawn). sd_login_monitor can now be used to watch
4533 VMs/containers coming and going.
4535 * .include is not allowed recursively anymore, and only in
4536 unit files. Usually it is better to use drop-in snippets in
4537 .d/*.conf anyway, as introduced with systemd 198.
4539 * systemd-analyze gained a new "critical-chain" command that
4540 determines the slowest chain of units run during system
4541 boot-up. It is very useful for tracking down where
4542 optimizing boot time is the most beneficial.
4544 * systemd will no longer allow manipulating service paths in
4545 the name=systemd:/system cgroup tree using ControlGroup= in
4546 units. (But is still fine with it in all other dirs.)
4548 * There's a new systemd-nspawn@.service service file that may
4549 be used to easily run nspawn containers as system
4550 services. With the container's root directory in
4551 /var/lib/container/foobar it is now sufficient to run
4552 "systemctl start systemd-nspawn@foobar.service" to boot it.
4554 * systemd-cgls gained a new parameter "--machine" to list only
4555 the processes within a certain container.
4557 * ConditionSecurity= now can check for "apparmor". We still
4558 are lacking checks for SMACK and IMA for this condition
4559 check though. Patches welcome!
4561 * A new configuration file /etc/systemd/sleep.conf has been
4562 added that may be used to configure which kernel operation
4563 systemd is supposed to execute when "suspend", "hibernate"
4564 or "hybrid-sleep" is requested. This makes the new kernel
4565 "freeze" state accessible to the user.
4567 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
4568 the passed argument if applicable.
4570 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
4571 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
4572 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
4573 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
4574 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
4575 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
4576 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
4581 * The output of 'systemctl list-jobs' got some polishing. The
4582 '--type=' argument may now be passed more than once. A new
4583 command 'systemctl list-sockets' has been added which shows
4584 a list of kernel sockets systemd is listening on with the
4585 socket units they belong to, plus the units these socket
4588 * The experimental libsystemd-bus library got substantial
4589 updates to work in conjunction with the (also experimental)
4590 kdbus kernel project. It works well enough to exchange
4591 messages with some sophistication. Note that kdbus is not
4592 ready yet, and the library is mostly an elaborate test case
4593 for now, and not installable.
4595 * systemd gained a new unit 'systemd-static-nodes.service'
4596 that generates static device nodes earlier during boot, and
4597 can run in conjunction with udev.
4599 * libsystemd-login gained a new call sd_pid_get_user_unit()
4600 to retrieve the user systemd unit a process is running
4601 in. This is useful for systems where systemd is used as
4604 * systemd-nspawn now places all containers in the new /machine
4605 top-level cgroup directory in the name=systemd
4606 hierarchy. libvirt will soon do the same, so that we get a
4607 uniform separation of /system, /user and /machine for system
4608 services, user processes and containers/virtual
4609 machines. This new cgroup hierarchy is also useful to stick
4610 stable names to specific container instances, which can be
4611 recognized later this way (this name may be controlled
4612 via systemd-nspawn's new -M switch). libsystemd-login also
4613 gained a new call sd_pid_get_machine_name() to retrieve the
4614 name of the container/VM a specific process belongs to.
4616 * bootchart can now store its data in the journal.
4618 * libsystemd-journal gained a new call
4619 sd_journal_add_conjunction() for AND expressions to the
4620 matching logic. This can be used to express more complex
4621 logical expressions.
4623 * journactl can now take multiple --unit= and --user-unit=
4626 * The cryptsetup logic now understands the "luks.key=" kernel
4627 command line switch for specifying a file to read the
4628 decryption key from. Also, if a configured key file is not
4629 found the tool will now automatically fall back to prompting
4632 * Python systemd.journal module was updated to wrap recently
4633 added functions from libsystemd-journal. The interface was
4634 changed to bring the low level interface in s.j._Reader
4635 closer to the C API, and the high level interface in
4636 s.j.Reader was updated to wrap and convert all data about
4639 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
4640 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
4641 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
4642 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
4643 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
4644 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
4648 * journalctl --update-catalog now understands a new --root=
4649 option to operate on catalogs found in a different root
4652 * During shutdown after systemd has terminated all running
4653 services a final killing loop kills all remaining left-over
4654 processes. We will now print the name of these processes
4655 when we send SIGKILL to them, since this usually indicates a
4658 * If /etc/crypttab refers to password files stored on
4659 configured mount points automatic dependencies will now be
4660 generated to ensure the specific mount is established first
4661 before the key file is attempted to be read.
4663 * 'systemctl status' will now show information about the
4664 network sockets a socket unit is listening on.
4666 * 'systemctl status' will also shown information about any
4667 drop-in configuration file for units. (Drop-In configuration
4668 files in this context are files such as
4669 /etc/systemd/systemd/foobar.service.d/*.conf)
4671 * systemd-cgtop now optionally shows summed up CPU times of
4672 cgroups. Press '%' while running cgtop to switch between
4673 percentage and absolute mode. This is useful to determine
4674 which cgroups use up the most CPU time over the entire
4675 runtime of the system. systemd-cgtop has also been updated
4676 to be 'pipeable' for processing with further shell tools.
4678 * 'hostnamectl set-hostname' will now allow setting of FQDN
4681 * The formatting and parsing of time span values has been
4682 changed. The parser now understands fractional expressions
4683 such as "5.5h". The formatter will now output fractional
4684 expressions for all time spans under 1min, i.e. "5.123456s"
4685 rather than "5s 123ms 456us". For time spans under 1s
4686 millisecond values are shown, for those under 1ms
4687 microsecond values are shown. This should greatly improve
4688 all time-related output of systemd.
4690 * libsystemd-login and libsystemd-journal gained new
4691 functions for querying the poll() events mask and poll()
4692 timeout value for integration into arbitrary event
4695 * localectl gained the ability to list available X11 keymaps
4696 (models, layouts, variants, options).
4698 * 'systemd-analyze dot' gained the ability to filter for
4699 specific units via shell-style globs, to create smaller,
4700 more useful graphs. I.e. it is now possible to create simple
4701 graphs of all the dependencies between only target units, or
4702 of all units that Avahi has dependencies with.
4704 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
4705 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
4706 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
4707 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
4708 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
4709 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
4710 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
4714 * The boot-time readahead implementation for rotating media
4715 will now read the read-ahead data in multiple passes which
4716 consist of all read requests made in equidistant time
4717 intervals. This means instead of strictly reading read-ahead
4718 data in its physical order on disk we now try to find a
4719 middle ground between physical and access time order.
4721 * /etc/os-release files gained a new BUILD_ID= field for usage
4722 on operating systems that provide continuous builds of OS
4725 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
4726 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
4727 William Douglas, Zbigniew Jędrzejewski-Szmek
4731 * systemd-python gained an API exposing libsystemd-daemon.
4733 * The SMACK setup logic gained support for uploading CIPSO
4736 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
4737 ReadOnlyDirectories= and InaccessibleDirectories= has
4738 changed. The private /tmp and /var/tmp directories are now
4739 shared by all processes of a service (which means
4740 ExecStartPre= may now leave data in /tmp that ExecStart= of
4741 the same service can still access). When a service is
4742 stopped its temporary directories are immediately deleted
4743 (normal clean-up with tmpfiles is still done in addition to
4746 * By default, systemd will now set a couple of sysctl
4747 variables in the kernel: the safe sysrq options are turned
4748 on, IP route verification is turned on, and source routing
4749 disabled. The recently added hardlink and softlink
4750 protection of the kernel is turned on. These settings should
4751 be reasonably safe, and good defaults for all new systems.
4753 * The predictable network naming logic may now be turned off
4754 with a new kernel command line switch: net.ifnames=0.
4756 * A new libsystemd-bus module has been added that implements a
4757 pretty complete D-Bus client library. For details see:
4759 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
4761 * journald will now explicitly flush the journal files to disk
4762 at the latest 5min after each write. The file will then also
4763 be marked offline until the next write. This should increase
4764 reliability in case of a crash. The synchronization delay
4765 can be configured via SyncIntervalSec= in journald.conf.
4767 * There's a new remote-fs-setup.target unit that can be used
4768 to pull in specific services when at least one remote file
4769 system is to be mounted.
4771 * There are new targets timers.target and paths.target as
4772 canonical targets to pull user timer and path units in
4773 from. This complements sockets.target with a similar
4774 purpose for socket units.
4776 * libudev gained a new call udev_device_set_attribute_value()
4777 to set sysfs attributes of a device.
4779 * The udev daemon now sets the default number of worker
4780 processes executed in parallel based on the number of available
4781 CPUs instead of the amount of available RAM. This is supposed
4782 to provide a more reliable default and limit a too aggressive
4783 paralellism for setups with 1000s of devices connected.
4785 Contributions from: Auke Kok, Colin Walters, Cristian
4786 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
4787 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
4788 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
4789 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
4790 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
4791 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
4792 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
4793 Zbigniew Jędrzejewski-Szmek
4797 * Configuration of unit files may now be extended via drop-in
4798 files without having to edit/override the unit files
4799 themselves. More specifically, if the administrator wants to
4800 change one value for a service file foobar.service he can
4801 now do so by dropping in a configuration snippet into
4802 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
4803 will load all these snippets and apply them on top of the
4804 main unit configuration file, possibly extending or
4805 overriding its settings. Using these drop-in snippets is
4806 generally nicer than the two earlier options for changing
4807 unit files locally: copying the files from
4808 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
4809 them there; or creating a new file in /etc/systemd/system/
4810 that incorporates the original one via ".include". Drop-in
4811 snippets into these .d/ directories can be placed in any
4812 directory systemd looks for units in, and the usual
4813 overriding semantics between /usr/lib, /etc and /run apply
4816 * Most unit file settings which take lists of items can now be
4817 reset by assigning the empty string to them. For example,
4818 normally, settings such as Environment=FOO=BAR append a new
4819 environment variable assignment to the environment block,
4820 each time they are used. By assigning Environment= the empty
4821 string the environment block can be reset to empty. This is
4822 particularly useful with the .d/*.conf drop-in snippets
4823 mentioned above, since this adds the ability to reset list
4824 settings from vendor unit files via these drop-ins.
4826 * systemctl gained a new "list-dependencies" command for
4827 listing the dependencies of a unit recursively.
4829 * Inhibitors are now honored and listed by "systemctl
4830 suspend", "systemctl poweroff" (and similar) too, not only
4831 GNOME. These commands will also list active sessions by
4834 * Resource limits (as exposed by the various control group
4835 controllers) can now be controlled dynamically at runtime
4836 for all units. More specifically, you can now use a command
4837 like "systemctl set-cgroup-attr foobar.service cpu.shares
4838 2000" to alter the CPU shares a specific service gets. These
4839 settings are stored persistently on disk, and thus allow the
4840 administrator to easily adjust the resource usage of
4841 services with a few simple commands. This dynamic resource
4842 management logic is also available to other programs via the
4843 bus. Almost any kernel cgroup attribute and controller is
4846 * systemd-vconsole-setup will now copy all font settings to
4847 all allocated VTs, where it previously applied them only to
4850 * libsystemd-login gained the new sd_session_get_tty() API
4853 * This release drops support for a few legacy or
4854 distribution-specific LSB facility names when parsing init
4855 scripts: $x-display-manager, $mail-transfer-agent,
4856 $mail-transport-agent, $mail-transfer-agent, $smtp,
4857 $null. Also, the mail-transfer-agent.target unit backing
4858 this has been removed. Distributions which want to retain
4859 compatibility with this should carry the burden for
4860 supporting this themselves and patch support for these back
4861 in, if they really need to. Also, the facilities $syslog and
4862 $local_fs are now ignored, since systemd does not support
4863 early-boot LSB init scripts anymore, and these facilities
4864 are implied anyway for normal services. syslog.target has
4867 * There are new bus calls on PID1's Manager object for
4868 cancelling jobs, and removing snapshot units. Previously,
4869 both calls were only available on the Job and Snapshot
4872 * systemd-journal-gatewayd gained SSL support.
4874 * The various "environment" files, such as /etc/locale.conf
4875 now support continuation lines with a backslash ("\") as
4876 last character in the line, similarly in style (but different)
4877 to how this is supported in shells.
4879 * For normal user processes the _SYSTEMD_USER_UNIT= field is
4880 now implicitly appended to every log entry logged. systemctl
4881 has been updated to filter by this field when operating on a
4882 user systemd instance.
4884 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
4885 CAP_AUDIT_CONTROL capabilities to the capabilities set for
4886 the container. This makes it easier to boot unmodified
4887 Fedora systems in a container, which however still requires
4888 audit=0 to be passed on the kernel command line. Auditing in
4889 kernel and userspace is unfortunately still too broken in
4890 context of containers, hence we recommend compiling it out
4891 of the kernel or using audit=0. Hopefully this will be fixed
4892 one day for good in the kernel.
4894 * nspawn gained the new --bind= and --bind-ro= parameters to
4895 bind mount specific directories from the host into the
4898 * nspawn will now mount its own devpts file system instance
4899 into the container, in order not to leak pty devices from
4900 the host into the container.
4902 * systemd will now read the firmware boot time performance
4903 information from the EFI variables, if the used boot loader
4904 supports this, and takes it into account for boot performance
4905 analysis via "systemd-analyze". This is currently supported
4906 only in conjunction with Gummiboot, but could be supported
4907 by other boot loaders too. For details see:
4909 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
4911 * A new generator has been added that automatically mounts the
4912 EFI System Partition (ESP) to /boot, if that directory
4913 exists, is empty, and no other file system has been
4914 configured to be mounted there.
4916 * logind will now send out PrepareForSleep(false) out
4917 unconditionally, after coming back from suspend. This may be
4918 used by applications as asynchronous notification for
4919 system resume events.
4921 * "systemctl unlock-sessions" has been added, that allows
4922 unlocking the screens of all user sessions at once, similar
4923 to how "systemctl lock-sessions" already locked all users
4924 sessions. This is backed by a new D-Bus call UnlockSessions().
4926 * "loginctl seat-status" will now show the master device of a
4927 seat. (i.e. the device of a seat that needs to be around for
4928 the seat to be considered available, usually the graphics
4931 * tmpfiles gained a new "X" line type, that allows
4932 configuration of files and directories (with wildcards) that
4933 shall be excluded from automatic cleanup ("aging").
4935 * udev default rules set the device node permissions now only
4936 at "add" events, and do not change them any longer with a
4937 later "change" event.
4939 * The log messages for lid events and power/sleep keypresses
4940 now carry a message ID.
4942 * We now have a substantially larger unit test suite, but this
4943 continues to be work in progress.
4945 * udevadm hwdb gained a new --root= parameter to change the
4946 root directory to operate relative to.
4948 * logind will now issue a background sync() request to the kernel
4949 early at shutdown, so that dirty buffers are flushed to disk early
4950 instead of at the last moment, in order to optimize shutdown
4953 * A new bootctl tool has been added that is an interface for
4954 certain boot loader operations. This is currently a preview
4955 and is likely to be extended into a small mechanism daemon
4956 like timedated, localed, hostnamed, and can be used by
4957 graphical UIs to enumerate available boot options, and
4958 request boot into firmware operations.
4960 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
4961 the rest of the package. It also has been updated to work
4962 correctly in initrds.
4964 * Policykit previously has been runtime optional, and is now
4965 also compile time optional via a configure switch.
4967 * systemd-analyze has been reimplemented in C. Also "systemctl
4968 dot" has moved into systemd-analyze.
4970 * "systemctl status" with no further parameters will now print
4971 the status of all active or failed units.
4973 * Operations such as "systemctl start" can now be executed
4974 with a new mode "--irreversible" which may be used to queue
4975 operations that cannot accidentally be reversed by a later
4976 job queuing. This is by default used to make shutdown
4977 requests more robust.
4979 * The Python API of systemd now gained a new module for
4980 reading journal files.
4982 * A new tool kernel-install has been added that can install
4983 kernel images according to the Boot Loader Specification:
4985 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
4987 * Boot time console output has been improved to provide
4988 animated boot time output for hanging jobs.
4990 * A new tool systemd-activate has been added which can be used
4991 to test socket activation with, directly from the command
4992 line. This should make it much easier to test and debug
4993 socket activation in daemons.
4995 * journalctl gained a new "--reverse" (or -r) option to show
4996 journal output in reverse order (i.e. newest line first).
4998 * journalctl gained a new "--pager-end" (or -e) option to jump
4999 to immediately jump to the end of the journal in the
5000 pager. This is only supported in conjunction with "less".
5002 * journalctl gained a new "--user-unit=" option, that works
5003 similarly to "--unit=" but filters for user units rather than
5006 * A number of unit files to ease adoption of systemd in
5007 initrds has been added. This moves some minimal logic from
5008 the various initrd implementations into systemd proper.
5010 * The journal files are now owned by a new group
5011 "systemd-journal", which exists specifically to allow access
5012 to the journal, and nothing else. Previously, we used the
5013 "adm" group for that, which however possibly covers more
5014 than just journal/log file access. This new group is now
5015 already used by systemd-journal-gatewayd to ensure this
5016 daemon gets access to the journal files and as little else
5017 as possible. Note that "make install" will also set FS ACLs
5018 up for /var/log/journal to give "adm" and "wheel" read
5019 access to it, in addition to "systemd-journal" which owns
5020 the journal files. We recommend that packaging scripts also
5021 add read access to "adm" + "wheel" to /var/log/journal, and
5022 all existing/future journal files. To normal users and
5023 administrators little changes, however packagers need to
5024 ensure to create the "systemd-journal" system group at
5025 package installation time.
5027 * The systemd-journal-gatewayd now runs as unprivileged user
5028 systemd-journal-gateway:systemd-journal-gateway. Packaging
5029 scripts need to create these system user/group at
5032 * timedated now exposes a new boolean property CanNTP that
5033 indicates whether a local NTP service is available or not.
5035 * systemd-detect-virt will now also detect xen PVs
5037 * The pstore file system is now mounted by default, if it is
5040 * In addition to the SELinux and IMA policies we will now also
5041 load SMACK policies at early boot.
5043 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
5044 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
5045 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
5046 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
5047 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
5048 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
5049 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
5050 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
5051 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
5052 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
5053 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
5054 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
5055 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
5056 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
5060 * Timer units now support calendar time events in addition to
5061 monotonic time events. That means you can now trigger a unit
5062 based on a calendar time specification such as "Thu,Fri
5063 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
5064 or fifth day of any month of the year 2013, given that it is
5065 a thursday or friday. This brings timer event support
5066 considerably closer to cron's capabilities. For details on
5067 the supported calendar time specification language see
5070 * udev now supports a number of different naming policies for
5071 network interfaces for predictable names, and a combination
5072 of these policies is now the default. Please see this wiki
5073 document for details:
5075 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
5077 * Auke Kok's bootchart implementation has been added to the
5078 systemd tree. It is an optional component that can graph the
5079 boot in quite some detail. It is one of the best bootchart
5080 implementations around and minimal in its code and
5083 * nss-myhostname has been integrated into the systemd source
5084 tree. nss-myhostname guarantees that the local hostname
5085 always stays resolvable via NSS. It has been a weak
5086 requirement of systemd-hostnamed since a long time, and
5087 since its code is actually trivial we decided to just
5088 include it in systemd's source tree. It can be turned off
5089 with a configure switch.
5091 * The read-ahead logic is now capable of properly detecting
5092 whether a btrfs file system is on SSD or rotating media, in
5093 order to optimize the read-ahead scheme. Previously, it was
5094 only capable of detecting this on traditional file systems
5097 * In udev, additional device properties are now read from the
5098 IAB in addition to the OUI database. Also, Bluetooth company
5099 identities are attached to the devices as well.
5101 * In service files %U may be used as specifier that is
5102 replaced by the configured user name of the service.
5104 * nspawn may now be invoked without a controlling TTY. This
5105 makes it suitable for invocation as its own service. This
5106 may be used to set up a simple containerized server system
5107 using only core OS tools.
5109 * systemd and nspawn can now accept socket file descriptors
5110 when they are started for socket activation. This enables
5111 implementation of socket activated nspawn
5112 containers. i.e. think about autospawning an entire OS image
5113 when the first SSH or HTTP connection is received. We expect
5114 that similar functionality will also be added to libvirt-lxc
5117 * journalctl will now suppress ANSI color codes when
5118 presenting log data.
5120 * systemctl will no longer show control group information for
5121 a unit if a the control group is empty anyway.
5123 * logind can now automatically suspend/hibernate/shutdown the
5126 * /etc/machine-info and hostnamed now also expose the chassis
5127 type of the system. This can be used to determine whether
5128 the local system is a laptop, desktop, handset or
5129 tablet. This information may either be configured by the
5130 user/vendor or is automatically determined from ACPI and DMI
5131 information if possible.
5133 * A number of PolicyKit actions are now bound together with
5134 "imply" rules. This should simplify creating UIs because
5135 many actions will now authenticate similar ones as well.
5137 * Unit files learnt a new condition ConditionACPower= which
5138 may be used to conditionalize a unit depending on whether an
5139 AC power source is connected or not, of whether the system
5140 is running on battery power.
5142 * systemctl gained a new "is-failed" verb that may be used in
5143 shell scripts and suchlike to check whether a specific unit
5144 is in the "failed" state.
5146 * The EnvironmentFile= setting in unit files now supports file
5147 globbing, and can hence be used to easily read a number of
5148 environment files at once.
5150 * systemd will no longer detect and recognize specific
5151 distributions. All distribution-specific #ifdeffery has been
5152 removed, systemd is now fully generic and
5153 distribution-agnostic. Effectively, not too much is lost as
5154 a lot of the code is still accessible via explicit configure
5155 switches. However, support for some distribution specific
5156 legacy configuration file formats has been dropped. We
5157 recommend distributions to simply adopt the configuration
5158 files everybody else uses now and convert the old
5159 configuration from packaging scripts. Most distributions
5160 already did that. If that's not possible or desirable,
5161 distributions are welcome to forward port the specific
5162 pieces of code locally from the git history.
5164 * When logging a message about a unit systemd will now always
5165 log the unit name in the message meta data.
5167 * localectl will now also discover system locale data that is
5168 not stored in locale archives, but directly unpacked.
5170 * logind will no longer unconditionally use framebuffer
5171 devices as seat masters, i.e. as devices that are required
5172 to be existing before a seat is considered preset. Instead,
5173 it will now look for all devices that are tagged as
5174 "seat-master" in udev. By default, framebuffer devices will
5175 be marked as such, but depending on local systems, other
5176 devices might be marked as well. This may be used to
5177 integrate graphics cards using closed source drivers (such
5178 as NVidia ones) more nicely into logind. Note however, that
5179 we recommend using the open source NVidia drivers instead,
5180 and no udev rules for the closed-source drivers will be
5181 shipped from us upstream.
5183 Contributions from: Adam Williamson, Alessandro Crismani, Auke
5184 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
5185 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
5186 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
5187 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
5188 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
5189 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
5190 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
5191 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
5192 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
5193 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
5194 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
5199 * udev gained support for loading additional device properties
5200 from an indexed database that is keyed by vendor/product IDs
5201 and similar device identifiers. For the beginning this
5202 "hwdb" is populated with data from the well-known PCI and
5203 USB database, but also includes PNP, ACPI and OID data. In
5204 the longer run this indexed database shall grow into
5205 becoming the one central database for non-essential
5206 userspace device metadata. Previously, data from the PCI/USB
5207 database was only attached to select devices, since the
5208 lookup was a relatively expensive operation due to O(n) time
5209 complexity (with n being the number of entries in the
5210 database). Since this is now O(1), we decided to add in this
5211 data for all devices where this is available, by
5212 default. Note that the indexed database needs to be rebuilt
5213 when new data files are installed. To achieve this you need
5214 to update your packaging scripts to invoke "udevadm hwdb
5215 --update" after installation of hwdb data files. For
5216 RPM-based distributions we introduced the new
5217 %udev_hwdb_update macro for this purpose.
5219 * The Journal gained support for the "Message Catalog", an
5220 indexed database to link up additional information with
5221 journal entries. For further details please check:
5223 http://www.freedesktop.org/wiki/Software/systemd/catalog
5225 The indexed message catalog database also needs to be
5226 rebuilt after installation of message catalog files. Use
5227 "journalctl --update-catalog" for this. For RPM-based
5228 distributions we introduced the %journal_catalog_update
5229 macro for this purpose.
5231 * The Python Journal bindings gained support for the standard
5232 Python logging framework.
5234 * The Journal API gained new functions for checking whether
5235 the underlying file system of a journal file is capable of
5236 properly reporting file change notifications, or whether
5237 applications that want to reflect journal changes "live"
5238 need to recheck journal files continuously in appropriate
5241 * It is now possible to set the "age" field for tmpfiles
5242 entries to 0, indicating that files matching this entry
5243 shall always be removed when the directories are cleaned up.
5245 * coredumpctl gained a new "gdb" verb which invokes gdb
5246 right-away on the selected coredump.
5248 * There's now support for "hybrid sleep" on kernels that
5249 support this, in addition to "suspend" and "hibernate". Use
5250 "systemctl hybrid-sleep" to make use of this.
5252 * logind's HandleSuspendKey= setting (and related settings)
5253 now gained support for a new "lock" setting to simply
5254 request the screen lock on all local sessions, instead of
5255 actually executing a suspend or hibernation.
5257 * systemd will now mount the EFI variables file system by
5260 * Socket units now gained support for configuration of the
5261 SMACK security label.
5263 * timedatectl will now output the time of the last and next
5264 daylight saving change.
5266 * We dropped support for various legacy and distro-specific
5267 concepts, such as insserv, early-boot SysV services
5268 (i.e. those for non-standard runlevels such as 'b' or 'S')
5269 or ArchLinux /etc/rc.conf support. We recommend the
5270 distributions who still need support this to either continue
5271 to maintain the necessary patches downstream, or find a
5272 different solution. (Talk to us if you have questions!)
5274 * Various systemd components will now bypass PolicyKit checks
5275 for root and otherwise handle properly if PolicyKit is not
5276 found to be around. This should fix most issues for
5277 PolicyKit-less systems. Quite frankly this should have been
5278 this way since day one. It is absolutely our intention to
5279 make systemd work fine on PolicyKit-less systems, and we
5280 consider it a bug if something does not work as it should if
5281 PolicyKit is not around.
5283 * For embedded systems it is now possible to build udev and
5284 systemd without blkid and/or kmod support.
5286 * "systemctl switch-root" is now capable of switching root
5287 more than once. I.e. in addition to transitions from the
5288 initrd to the host OS it is now possible to transition to
5289 further OS images from the host. This is useful to implement
5290 offline updating tools.
5292 * Various other additions have been made to the RPM macros
5293 shipped with systemd. Use %udev_rules_update() after
5294 installing new udev rules files. %_udevhwdbdir,
5295 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
5296 %_sysctldir are now available which resolve to the right
5297 directories for packages to place various data files in.
5299 * journalctl gained the new --full switch (in addition to
5300 --all, to disable ellipsation for long messages.
5302 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
5303 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
5304 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
5305 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
5306 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
5307 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
5308 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
5309 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
5310 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
5314 * journalctl gained new --since= and --until= switches to
5315 filter by time. It also now supports nice filtering for
5316 units via --unit=/-u.
5318 * Type=oneshot services may use ExecReload= and do the
5321 * The journal daemon now supports time-based rotation and
5322 vacuuming, in addition to the usual disk-space based
5325 * The journal will now index the available field values for
5326 each field name. This enables clients to show pretty drop
5327 downs of available match values when filtering. The bash
5328 completion of journalctl has been updated
5329 accordingly. journalctl gained a new switch -F to list all
5330 values a certain field takes in the journal database.
5332 * More service events are now written as structured messages
5333 to the journal, and made recognizable via message IDs.
5335 * The timedated, localed and hostnamed mini-services which
5336 previously only provided support for changing time, locale
5337 and hostname settings from graphical DEs such as GNOME now
5338 also have a minimal (but very useful) text-based client
5339 utility each. This is probably the nicest way to changing
5340 these settings from the command line now, especially since
5341 it lists available options and is fully integrated with bash
5344 * There's now a new tool "systemd-coredumpctl" to list and
5345 extract coredumps from the journal.
5347 * We now install a README each in /var/log/ and
5348 /etc/rc.d/init.d explaining where the system logs and init
5349 scripts went. This hopefully should help folks who go to
5350 that dirs and look into the otherwise now empty void and
5351 scratch their heads.
5353 * When user-services are invoked (by systemd --user) the
5354 $MANAGERPID env var is set to the PID of systemd.
5356 * SIGRTMIN+24 when sent to a --user instance will now result
5357 in immediate termination of systemd.
5359 * gatewayd received numerous feature additions such as a
5360 "follow" mode, for live syncing and filtering.
5362 * browse.html now allows filtering and showing detailed
5363 information on specific entries. Keyboard navigation and
5364 mouse screen support has been added.
5366 * gatewayd/journalctl now supports HTML5/JSON
5367 Server-Sent-Events as output.
5369 * The SysV init script compatibility logic will now
5370 heuristically determine whether a script supports the
5371 "reload" verb, and only then make this available as
5374 * "systemctl status --follow" has been removed, use "journalctl
5377 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
5378 have been removed since they are hardly useful to be
5381 * And I'd like to take the opportunity to specifically mention
5382 Zbigniew for his great contributions. Zbigniew, you rock!
5384 Contributions from: Andrew Eikum, Christian Hesse, Colin
5385 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
5386 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
5387 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
5388 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
5389 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
5390 Jędrzejewski-Szmek, Сковорода Никита Андреевич
5394 * If /etc/vconsole.conf is non-existent or empty we will no
5395 longer load any console font or key map at boot by
5396 default. Instead the kernel defaults will be left
5397 intact. This is definitely the right thing to do, as no
5398 configuration should mean no configuration, and hard-coding
5399 font names that are different on all archs is probably a bad
5400 idea. Also, the kernel default key map and font should be
5401 good enough for most cases anyway, and mostly identical to
5402 the userspace fonts/key maps we previously overloaded them
5403 with. If distributions want to continue to default to a
5404 non-kernel font or key map they should ship a default
5405 /etc/vconsole.conf with the appropriate contents.
5407 Contributions from: Colin Walters, Daniel J Walsh, Dave
5408 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
5409 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
5413 * journalctl gained a new --cursor= switch to show entries
5414 starting from the specified location in the journal.
5416 * We now enforce a size limit on journal entry fields exported
5417 with "-o json" in journalctl. Fields larger than 4K will be
5418 assigned null. This can be turned off with --all.
5420 * An (optional) journal gateway daemon is now available as
5421 "systemd-journal-gatewayd.service". This service provides
5422 access to the journal via HTTP and JSON. This functionality
5423 will be used to implement live log synchronization in both
5424 pull and push modes, but has various other users too, such
5425 as easy log access for debugging of embedded devices. Right
5426 now it is already useful to retrieve the journal via HTTP:
5428 # systemctl start systemd-journal-gatewayd.service
5429 # wget http://localhost:19531/entries
5431 This will download the journal contents in a
5432 /var/log/messages compatible format. The same as JSON:
5434 # curl -H"Accept: application/json" http://localhost:19531/entries
5436 This service is also accessible via a web browser where a
5437 single static HTML5 app is served that uses the JSON logic
5438 to enable the user to do some basic browsing of the
5439 journal. This will be extended later on. Here's an example
5440 screenshot of this app in its current state:
5442 http://0pointer.de/public/journal-gatewayd
5444 Contributions from: Kay Sievers, Lennart Poettering, Robert
5445 Milasan, Tom Gundersen
5449 * The bash completion logic is now available for journalctl
5452 * We do not mount the "cpuset" controller anymore together with
5453 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
5454 started if no parameters are assigned to it. "cpuset" hence
5455 broke code that assumed it could create "cpu" groups and
5458 * journalctl -f will now subscribe to terminal size changes,
5459 and line break accordingly.
5461 Contributions from: Dave Reisner, Kay Sievers, Lennart
5462 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
5466 * nspawn will now create a symlink /etc/localtime in the
5467 container environment, copying the host's timezone
5468 setting. Previously this has been done via a bind mount, but
5469 since symlinks cannot be bind mounted this has now been
5470 changed to create/update the appropriate symlink.
5472 * journalctl -n's line number argument is now optional, and
5473 will default to 10 if omitted.
5475 * journald will now log the maximum size the journal files may
5476 take up on disk. This is particularly useful if the default
5477 built-in logic of determining this parameter from the file
5478 system size is used. Use "systemctl status
5479 systemd-journald.service" to see this information.
5481 * The multi-seat X wrapper tool has been stripped down. As X
5482 is now capable of enumerating graphics devices via udev in a
5483 seat-aware way the wrapper is not strictly necessary
5484 anymore. A stripped down temporary stop-gap is still shipped
5485 until the upstream display managers have been updated to
5486 fully support the new X logic. Expect this wrapper to be
5487 removed entirely in one of the next releases.
5489 * HandleSleepKey= in logind.conf has been split up into
5490 HandleSuspendKey= and HandleHibernateKey=. The old setting
5491 is not available anymore. X11 and the kernel are
5492 distinguishing between these keys and we should too. This
5493 also means the inhibition lock for these keys has been split
5496 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
5497 Poettering, Lukas Nykryn, Václav Pavlín
5501 * Whenever a unit changes state we will now log this to the
5502 journal and show along the unit's own log output in
5505 * ConditionPathIsMountPoint= can now properly detect bind
5506 mount points too. (Previously, a bind mount of one file
5507 system to another place in the same file system could not be
5508 detected as mount, since they shared struct stat's st_dev
5511 * We will now mount the cgroup controllers cpu, cpuacct,
5512 cpuset and the controllers net_cls, net_prio together by
5515 * nspawn containers will now have a virtualized boot
5516 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
5517 over with a randomized ID at container initialization). This
5518 has the effect of making "journalctl -b" do the right thing
5521 * The JSON output journal serialization has been updated not
5522 to generate "endless" list objects anymore, but rather one
5523 JSON object per line. This is more in line how most JSON
5524 parsers expect JSON objects. The new output mode
5525 "json-pretty" has been added to provide similar output, but
5526 neatly aligned for readability by humans.
5528 * We dropped all explicit sync() invocations in the shutdown
5529 code. The kernel does this implicitly anyway in the kernel
5530 reboot() syscall. halt(8)'s -n option is now a compatibility
5533 * We now support virtualized reboot() in containers, as
5534 supported by newer kernels. We will fall back to exit() if
5535 CAP_SYS_REBOOT is not available to the container. Also,
5536 nspawn makes use of this now and will actually reboot the
5537 container if the containerized OS asks for that.
5539 * journalctl will only show local log output by default
5540 now. Use --merge (-m) to show remote log output, too.
5542 * libsystemd-journal gained the new sd_journal_get_usage()
5543 call to determine the current disk usage of all journal
5544 files. This is exposed in the new "journalctl --disk-usage"
5547 * journald gained a new configuration setting SplitMode= in
5548 journald.conf which may be used to control how user journals
5549 are split off. See journald.conf(5) for details.
5551 * A new condition type ConditionFileNotEmpty= has been added.
5553 * tmpfiles' "w" lines now support file globbing, to write
5554 multiple files at once.
5556 * We added Python bindings for the journal submission
5557 APIs. More Python APIs for a number of selected APIs will
5558 likely follow. Note that we intend to add native bindings
5559 only for the Python language, as we consider it common
5560 enough to deserve bindings shipped within systemd. There are
5561 various projects outside of systemd that provide bindings
5562 for languages such as PHP or Lua.
5564 * Many conditions will now resolve specifiers such as %i. In
5565 addition, PathChanged= and related directives of .path units
5566 now support specifiers as well.
5568 * There's now a new RPM macro definition for the system preset
5571 * journald will now warn if it ca not forward a message to the
5572 syslog daemon because its socket is full.
5574 * timedated will no longer write or process /etc/timezone,
5575 except on Debian. As we do not support late mounted /usr
5576 anymore /etc/localtime always being a symlink is now safe,
5577 and hence the information in /etc/timezone is not necessary
5580 * logind will now always reserve one VT for a text getty (VT6
5581 by default). Previously if more than 6 X sessions where
5582 started they took up all the VTs with auto-spawned gettys,
5583 so that no text gettys were available anymore.
5585 * udev will now automatically inform the btrfs kernel logic
5586 about btrfs RAID components showing up. This should make
5587 simple hotplug based btrfs RAID assembly work.
5589 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
5590 (but not for its children which will stay at the kernel
5591 default). This should allow setups with a lot more listening
5594 * systemd will now always pass the configured timezone to the
5595 kernel at boot. timedated will do the same when the timezone
5598 * logind's inhibition logic has been updated. By default,
5599 logind will now handle the lid switch, the power and sleep
5600 keys all the time, even in graphical sessions. If DEs want
5601 to handle these events on their own they should take the new
5602 handle-power-key, handle-sleep-key and handle-lid-switch
5603 inhibitors during their runtime. A simple way to achieve
5604 that is to invoke the DE wrapped in an invocation of:
5606 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
5608 * Access to unit operations is now checked via SELinux taking
5609 the unit file label and client process label into account.
5611 * systemd will now notify the administrator in the journal
5612 when he over-mounts a non-empty directory.
5614 * There are new specifiers that are resolved in unit files,
5615 for the host name (%H), the machine ID (%m) and the boot ID
5618 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
5619 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
5620 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
5621 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
5622 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
5623 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5624 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
5628 * Support for reading structured kernel messages from
5629 /dev/kmsg has now been added and is enabled by default.
5631 * Support for reading kernel messages from /proc/kmsg has now
5632 been removed. If you want kernel messages in the journal
5633 make sure to run a recent kernel (>= 3.5) that supports
5634 reading structured messages from /dev/kmsg (see
5635 above). /proc/kmsg is now exclusive property of classic
5636 syslog daemons again.
5638 * The libudev API gained the new
5639 udev_device_new_from_device_id() call.
5641 * The logic for file system namespace (ReadOnlyDirectory=,
5642 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
5643 require pivot_root() anymore. This means fewer temporary
5644 directories are created below /tmp for this feature.
5646 * nspawn containers will now see and receive all submounts
5647 made on the host OS below the root file system of the
5650 * Forward Secure Sealing is now supported for Journal files,
5651 which provide cryptographical sealing of journal files so
5652 that attackers cannot alter log history anymore without this
5653 being detectable. Lennart will soon post a blog story about
5654 this explaining it in more detail.
5656 * There are two new service settings RestartPreventExitStatus=
5657 and SuccessExitStatus= which allow configuration of exit
5658 status (exit code or signal) which will be excepted from the
5659 restart logic, resp. consider successful.
5661 * journalctl gained the new --verify switch that can be used
5662 to check the integrity of the structure of journal files and
5663 (if Forward Secure Sealing is enabled) the contents of
5666 * nspawn containers will now be run with /dev/stdin, /dev/fd/
5667 and similar symlinks pre-created. This makes running shells
5668 as container init process a lot more fun.
5670 * The fstab support can now handle PARTUUID= and PARTLABEL=
5673 * A new ConditionHost= condition has been added to match
5674 against the hostname (with globs) and machine ID. This is
5675 useful for clusters where a single OS image is used to
5676 provision a large number of hosts which shall run slightly
5677 different sets of services.
5679 * Services which hit the restart limit will now be placed in a
5682 Contributions from: Bertram Poettering, Dave Reisner, Huang
5683 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
5684 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
5688 * When running in --user mode systemd will now become a
5689 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
5690 tree a lot more organized.
5692 * A new PartOf= unit dependency type has been introduced that
5693 may be used to group services in a natural way.
5695 * "systemctl enable" may now be used to enable instances of
5698 * journalctl now prints error log levels in red, and
5699 warning/notice log levels in bright white. It also supports
5700 filtering by log level now.
5702 * cgtop gained a new -n switch (similar to top), to configure
5703 the maximum number of iterations to run for. It also gained
5704 -b, to run in batch mode (accepting no input).
5706 * The suffix ".service" may now be omitted on most systemctl
5707 command lines involving service unit names.
5709 * There's a new bus call in logind to lock all sessions, as
5710 well as a loginctl verb for it "lock-sessions".
5712 * libsystemd-logind.so gained a new call sd_journal_perror()
5713 that works similar to libc perror() but logs to the journal
5714 and encodes structured information about the error number.
5716 * /etc/crypttab entries now understand the new keyfile-size=
5719 * shutdown(8) now can send a (configurable) wall message when
5720 a shutdown is cancelled.
5722 * The mount propagation mode for the root file system will now
5723 default to "shared", which is useful to make containers work
5724 nicely out-of-the-box so that they receive new mounts from
5725 the host. This can be undone locally by running "mount
5726 --make-rprivate /" if needed.
5728 * The prefdm.service file has been removed. Distributions
5729 should maintain this unit downstream if they intend to keep
5730 it around. However, we recommend writing normal unit files
5731 for display managers instead.
5733 * Since systemd is a crucial part of the OS we will now
5734 default to a number of compiler switches that improve
5735 security (hardening) such as read-only relocations, stack
5736 protection, and suchlike.
5738 * The TimeoutSec= setting for services is now split into
5739 TimeoutStartSec= and TimeoutStopSec= to allow configuration
5740 of individual time outs for the start and the stop phase of
5743 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
5744 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
5745 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
5746 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
5747 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
5748 Gundersen, Zbigniew Jędrzejewski-Szmek
5752 * The journal and id128 C APIs are now fully documented as man
5755 * Extra safety checks have been added when transitioning from
5756 the initial RAM disk to the main system to avoid accidental
5759 * /etc/crypttab entries now understand the new keyfile-offset=
5762 * systemctl -t can now be used to filter by unit load state.
5764 * The journal C API gained the new sd_journal_wait() call to
5765 make writing synchronous journal clients easier.
5767 * journalctl gained the new -D switch to show journals from a
5770 * journalctl now displays a special marker between log
5771 messages of two different boots.
5773 * The journal is now explicitly flushed to /var via a service
5774 systemd-journal-flush.service, rather than implicitly simply
5775 by seeing /var/log/journal to be writable.
5777 * journalctl (and the journal C APIs) can now match for much
5778 more complex expressions, with alternatives and
5781 * When transitioning from the initial RAM disk to the main
5782 system we will now kill all processes in a killing spree to
5783 ensure no processes stay around by accident.
5785 * Three new specifiers may be used in unit files: %u, %h, %s
5786 resolve to the user name, user home directory resp. user
5787 shell. This is useful for running systemd user instances.
5789 * We now automatically rotate journal files if their data
5790 object hash table gets a fill level > 75%. We also size the
5791 hash table based on the configured maximum file size. This
5792 together should lower hash collisions drastically and thus
5793 speed things up a bit.
5795 * journalctl gained the new "--header" switch to introspect
5796 header data of journal files.
5798 * A new setting SystemCallFilters= has been added to services
5799 which may be used to apply blacklists or whitelists to
5800 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
5802 * nspawn gained a new --link-journal= switch (and quicker: -j)
5803 to link the container journal with the host. This makes it
5804 very easy to centralize log viewing on the host for all
5805 guests while still keeping the journal files separated.
5807 * Many bugfixes and optimizations
5809 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
5810 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
5811 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
5816 * Several tools now understand kernel command line arguments,
5817 which are only read when run in an initial RAM disk. They
5818 usually follow closely their normal counterparts, but are
5821 * There's a new tool to analyze the readahead files that are
5822 automatically generated at boot. Use:
5824 /usr/lib/systemd/systemd-readahead analyze /.readahead
5826 * We now provide an early debug shell on tty9 if this enabled. Use:
5828 systemctl enable debug-shell.service
5830 * All plymouth related units have been moved into the Plymouth
5831 package. Please make sure to upgrade your Plymouth version
5834 * systemd-tmpfiles now supports getting passed the basename of
5835 a configuration file only, in which case it will look for it
5836 in all appropriate directories automatically.
5838 * udevadm info now takes a /dev or /sys path as argument, and
5839 does the right thing. Example:
5841 udevadm info /dev/sda
5842 udevadm info /sys/class/block/sda
5844 * systemctl now prints a warning if a unit is stopped but a
5845 unit that might trigger it continues to run. Example: a
5846 service is stopped but the socket that activates it is left
5849 * "systemctl status" will now mention if the log output was
5850 shortened due to rotation since a service has been started.
5852 * The journal API now exposes functions to determine the
5853 "cutoff" times due to rotation.
5855 * journald now understands SIGUSR1 and SIGUSR2 for triggering
5856 immediately flushing of runtime logs to /var if possible,
5857 resp. for triggering immediate rotation of the journal
5860 * It is now considered an error if a service is attempted to
5861 be stopped that is not loaded.
5863 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
5865 * systemd-analyze now supports Python 3
5867 * tmpfiles now supports cleaning up directories via aging
5868 where the first level dirs are always kept around but
5869 directories beneath it automatically aged. This is enabled
5870 by prefixing the age field with '~'.
5872 * Seat objects now expose CanGraphical, CanTTY properties
5873 which is required to deal with very fast bootups where the
5874 display manager might be running before the graphics drivers
5875 completed initialization.
5877 * Seat objects now expose a State property.
5879 * We now include RPM macros for service enabling/disabling
5880 based on the preset logic. We recommend RPM based
5881 distributions to make use of these macros if possible. This
5882 makes it simpler to reuse RPM spec files across
5885 * We now make sure that the collected systemd unit name is
5886 always valid when services log to the journal via
5889 * There's a new man page kernel-command-line(7) detailing all
5890 command line options we understand.
5892 * The fstab generator may now be disabled at boot by passing
5893 fstab=0 on the kernel command line.
5895 * A new kernel command line option modules-load= is now understood
5896 to load a specific kernel module statically, early at boot.
5898 * Unit names specified on the systemctl command line are now
5899 automatically escaped as needed. Also, if file system or
5900 device paths are specified they are automatically turned
5901 into the appropriate mount or device unit names. Example:
5903 systemctl status /home
5904 systemctl status /dev/sda
5906 * The SysVConsole= configuration option has been removed from
5907 system.conf parsing.
5909 * The SysV search path is no longer exported on the D-Bus
5912 * The Names= option is been removed from unit file parsing.
5914 * There's a new man page bootup(7) detailing the boot process.
5916 * Every unit and every generator we ship with systemd now
5917 comes with full documentation. The self-explanatory boot is
5920 * A couple of services gained "systemd-" prefixes in their
5921 name if they wrap systemd code, rather than only external
5922 code. Among them fsck@.service which is now
5923 systemd-fsck@.service.
5925 * The HaveWatchdog property has been removed from the D-Bus
5928 * systemd.confirm_spawn= on the kernel command line should now
5931 * There's a new man page crypttab(5) which details all options
5932 we actually understand.
5934 * systemd-nspawn gained a new --capability= switch to pass
5935 additional capabilities to the container.
5937 * timedated will now read known NTP implementation unit names
5938 from /usr/lib/systemd/ntp-units.d/*.list,
5939 systemd-timedated-ntp.target has been removed.
5941 * journalctl gained a new switch "-b" that lists log data of
5942 the current boot only.
5944 * The notify socket is in the abstract namespace again, in
5945 order to support daemons which chroot() at start-up.
5947 * There is a new Storage= configuration option for journald
5948 which allows configuration of where log data should go. This
5949 also provides a way to disable journal logging entirely, so
5950 that data collected is only forwarded to the console, the
5951 kernel log buffer or another syslog implementation.
5953 * Many bugfixes and optimizations
5955 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
5956 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
5957 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
5958 Shawn Landden, Tom Gundersen
5962 * "systemctl help <unit>" now shows the man page if one is
5965 * Several new man pages have been added.
5967 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
5968 MaxLevelConsole= can now be specified in
5969 journald.conf. These options allow reducing the amount of
5970 data stored on disk or forwarded by the log level.
5972 * TimerSlackNSec= can now be specified in system.conf for
5973 PID1. This allows system-wide power savings.
5975 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
5976 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
5981 * logind is now capable of (optionally) handling power and
5982 sleep keys as well as the lid switch.
5984 * journalctl now understands the syntax "journalctl
5985 /usr/bin/avahi-daemon" to get all log output of a specific
5988 * CapabilityBoundingSet= in system.conf now also influences
5989 the capability bound set of usermode helpers of the kernel.
5991 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
5992 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
5993 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
5994 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
5998 * Note that we skipped 139 releases here in order to set the
5999 new version to something that is greater than both udev's
6000 and systemd's most recent version number.
6002 * udev: all udev sources are merged into the systemd source tree now.
6003 All future udev development will happen in the systemd tree. It
6004 is still fully supported to use the udev daemon and tools without
6005 systemd running, like in initramfs or other init systems. Building
6006 udev though, will require the *build* of the systemd tree, but
6007 udev can be properly *run* without systemd.
6009 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
6010 should be used to create dead device nodes as workarounds for broken
6013 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
6014 no longer supported. udev_monitor_new_from_netlink() needs to be
6015 used to subscribe to events.
6017 * udev: when udevd is started by systemd, processes which are left
6018 behind by forking them off of udev rules, are unconditionally cleaned
6019 up and killed now after the event handling has finished. Services or
6020 daemons must be started as systemd services. Services can be
6021 pulled-in by udev to get started, but they can no longer be directly
6022 forked by udev rules.
6024 * udev: the daemon binary is called systemd-udevd now and installed
6025 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
6026 to adapt to that, create symlink, or rename the binary after building
6029 * libudev no longer provides these symbols:
6030 udev_monitor_from_socket()
6031 udev_queue_get_failed_list_entry()
6032 udev_get_{dev,sys,run}_path()
6033 The versions number was bumped and symbol versioning introduced.
6035 * systemd-loginctl and systemd-journalctl have been renamed
6036 to loginctl and journalctl to match systemctl.
6038 * The config files: /etc/systemd/systemd-logind.conf and
6039 /etc/systemd/systemd-journald.conf have been renamed to
6040 logind.conf and journald.conf. Package updates should rename
6041 the files to the new names on upgrade.
6043 * For almost all files the license is now LGPL2.1+, changed
6044 from the previous GPL2.0+. Exceptions are some minor stuff
6045 of udev (which will be changed to LGPL2.1 eventually, too),
6046 and the MIT licensed sd-daemon.[ch] library that is suitable
6047 to be used as drop-in files.
6049 * systemd and logind now handle system sleep states, in
6050 particular suspending and hibernating.
6052 * logind now implements a sleep/shutdown/idle inhibiting logic
6053 suitable for a variety of uses. Soonishly Lennart will blog
6054 about this in more detail.
6056 * var-run.mount and var-lock.mount are no longer provided
6057 (which prevously bind mounted these directories to their new
6058 places). Distributions which have not converted these
6059 directories to symlinks should consider stealing these files
6060 from git history and add them downstream.
6062 * We introduced the Documentation= field for units and added
6063 this to all our shipped units. This is useful to make it
6064 easier to explore the boot and the purpose of the various
6067 * All smaller setup units (such as
6068 systemd-vconsole-setup.service) now detect properly if they
6069 are run in a container and are skipped when
6070 appropriate. This guarantees an entirely noise-free boot in
6071 Linux container environments such as systemd-nspawn.
6073 * A framework for implementing offline system updates is now
6074 integrated, for details see:
6075 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
6077 * A new service type Type=idle is available now which helps us
6078 avoiding ugly interleaving of getty output and boot status
6081 * There's now a system-wide CapabilityBoundingSet= option to
6082 globally reduce the set of capabilities for the
6083 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
6084 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
6085 even CAP_NET_ADMIN system-wide for secure systems.
6087 * There are now system-wide DefaultLimitXXX= options to
6088 globally change the defaults of the various resource limits
6089 for all units started by PID 1.
6091 * Harald Hoyer's systemd test suite has been integrated into
6092 systemd which allows easy testing of systemd builds in qemu
6093 and nspawn. (This is really awesome! Ask us for details!)
6095 * The fstab parser is now implemented as generator, not inside
6098 * systemctl will now warn you if .mount units generated from
6099 /etc/fstab are out of date due to changes in fstab that
6100 have not been read by systemd yet.
6102 * systemd is now suitable for usage in initrds. Dracut has
6103 already been updated to make use of this. With this in place
6104 initrds get a slight bit faster but primarily are much
6105 easier to introspect and debug since "systemctl status" in
6106 the host system can be used to introspect initrd services,
6107 and the journal from the initrd is kept around too.
6109 * systemd-delta has been added, a tool to explore differences
6110 between user/admin configuration and vendor defaults.
6112 * PrivateTmp= now affects both /tmp and /var/tmp.
6114 * Boot time status messages are now much prettier and feature
6115 proper english language. Booting up systemd has never been
6118 * Read-ahead pack files now include the inode number of all
6119 files to pre-cache. When the inode changes the pre-caching
6120 is not attempted. This should be nicer to deal with updated
6121 packages which might result in changes of read-ahead
6124 * We now temporaritly lower the kernel's read_ahead_kb variable
6125 when collecting read-ahead data to ensure the kernel's
6126 built-in read-ahead does not add noise to our measurements
6127 of necessary blocks to pre-cache.
6129 * There's now RequiresMountsFor= to add automatic dependencies
6130 for all mounts necessary for a specific file system path.
6132 * MountAuto= and SwapAuto= have been removed from
6133 system.conf. Mounting file systems at boot has to take place
6136 * nspawn now learned a new switch --uuid= to set the machine
6137 ID on the command line.
6139 * nspawn now learned the -b switch to automatically search
6142 * vt102 is now the default TERM for serial TTYs, upgraded from
6145 * systemd-logind now works on VT-less systems.
6147 * The build tree has been reorganized. The individual
6148 components now have directories of their own.
6150 * A new condition type ConditionPathIsReadWrite= is now available.
6152 * nspawn learned the new -C switch to create cgroups for the
6153 container in other hierarchies.
6155 * We now have support for hardware watchdogs, configurable in
6158 * The scheduled shutdown logic now has a public API.
6160 * We now mount /tmp as tmpfs by default, but this can be
6161 masked and /etc/fstab can override it.
6163 * Since udisks does not make use of /media anymore we are not
6164 mounting a tmpfs on it anymore.
6166 * journalctl gained a new --local switch to only interleave
6167 locally generated journal files.
6169 * We can now load the IMA policy at boot automatically.
6171 * The GTK tools have been split off into a systemd-ui.
6173 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
6174 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
6175 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
6176 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
6177 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
6178 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
6179 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
6180 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
6181 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
6186 * This is mostly a bugfix release
6188 * Support optional initialization of the machine ID from the
6189 KVM or container configured UUID.
6191 * Support immediate reboots with "systemctl reboot -ff"
6193 * Show /etc/os-release data in systemd-analyze output
6195 * Many bugfixes for the journal, including endianness fixes and
6196 ensuring that disk space enforcement works
6198 * sd-login.h is C++ comptaible again
6200 * Extend the /etc/os-release format on request of the Debian
6203 * We now refuse non-UTF8 strings used in various configuration
6204 and unit files. This is done to ensure we do not pass invalid
6205 data over D-Bus or expose it elsewhere.
6207 * Register Mimo USB Screens as suitable for automatic seat
6210 * Read SELinux client context from journal clients in a race
6213 * Reorder configuration file lookup order. /etc now always
6214 overrides /run in order to allow the administrator to always
6215 and unconditionally override vendor-supplied or
6216 automatically generated data.
6218 * The various user visible bits of the journal now have man
6219 pages. We still lack man pages for the journal API calls
6222 * We now ship all man pages in HTML format again in the
6225 Contributions from: Dave Reisner, Dirk Eibach, Frederic
6226 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
6227 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
6232 * This is mostly a bugfix release
6234 * systems lacking /etc/os-release are no longer supported.
6236 * Various functionality updates to libsystemd-login.so
6238 * Track class of PAM logins to distinguish greeters from
6241 Contributions from: Kay Sievers, Lennart Poettering, Michael
6246 * This is an important bugfix release for v41.
6248 * Building man pages is now optional which should be useful
6249 for those building systemd from git but unwilling to install
6252 * Watchdog support for supervising services is now usable. In
6253 a future release support for hardware watchdogs
6254 (i.e. /dev/watchdog) will be added building on this.
6256 * Service start rate limiting is now configurable and can be
6257 turned off per service. When a start rate limit is hit a
6258 reboot can automatically be triggered.
6260 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
6262 Contributions from: Benjamin Franzke, Bill Nottingham,
6263 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
6264 Schmidt, Michał Górny, Piotr Drąg
6268 * The systemd binary is installed /usr/lib/systemd/systemd now;
6269 An existing /sbin/init symlink needs to be adapted with the
6272 * The code that loads kernel modules has been ported to invoke
6273 libkmod directly, instead of modprobe. This means we do not
6274 support systems with module-init-tools anymore.
6276 * Watchdog support is now already useful, but still not
6279 * A new kernel command line option systemd.setenv= is
6280 understood to set system wide environment variables
6281 dynamically at boot.
6283 * We now limit the set of capabilities of systemd-journald.
6285 * We now set SIGPIPE to ignore by default, since it only is
6286 useful in shell pipelines, and has little use in general
6287 code. This can be disabled with IgnoreSIPIPE=no in unit
6290 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
6291 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
6296 * This is mostly a bugfix release
6298 * We now expose the reason why a service failed in the
6299 "Result" D-Bus property.
6301 * Rudimentary service watchdog support (will be completed over
6302 the next few releases.)
6304 * When systemd forks off in order execute some service we will
6305 now immediately changes its argv[0] to reflect which process
6306 it will execute. This is useful to minimize the time window
6307 with a generic argv[0], which makes bootcharts more useful
6309 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
6310 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
6311 Mike Kazantsev, Ray Strode
6315 * This is mostly a test release, but incorporates many
6318 * New systemd-cgtop tool to show control groups by their
6321 * Linking against libacl for ACLs is optional again. If
6322 disabled, support tracking device access for active logins
6323 goes becomes unavailable, and so does access to the user
6324 journals by the respective users.
6326 * If a group "adm" exists, journal files are automatically
6327 owned by them, thus allow members of this group full access
6328 to the system journal as well as all user journals.
6330 * The journal now stores the SELinux context of the logging
6331 client for all entries.
6333 * Add C++ inclusion guards to all public headers
6335 * New output mode "cat" in the journal to print only text
6336 messages, without any meta data like date or time.
6338 * Include tiny X server wrapper as a temporary stop-gap to
6339 teach XOrg udev display enumeration. This is used by display
6340 managers such as gdm, and will go away as soon as XOrg
6341 learned native udev hotplugging for display devices.
6343 * Add new systemd-cat tool for executing arbitrary programs
6344 with STDERR/STDOUT connected to the journal. Can also act as
6345 BSD logger replacement, and does so by default.
6347 * Optionally store all locally generated coredumps in the
6348 journal along with meta data.
6350 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
6351 writing short strings to files (for usage for /sys), and for
6352 creating symlinks, character and block device nodes.
6354 * New unit file option ControlGroupPersistent= to make cgroups
6355 persistent, following the mechanisms outlined in
6356 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
6358 * Support multiple local RTCs in a sane way
6360 * No longer monopolize IO when replaying readahead data on
6361 rotating disks, since we might starve non-file-system IO to
6362 death, since fanotify() will not see accesses done by blkid,
6365 * Do not show kernel threads in systemd-cgls anymore, unless
6366 requested with new -k switch.
6368 Contributions from: Dan Horák, Kay Sievers, Lennart
6369 Poettering, Michal Schmidt
6373 * This is mostly a test release, but incorporates many
6376 * The git repository moved to:
6377 git://anongit.freedesktop.org/systemd/systemd
6378 ssh://git.freedesktop.org/git/systemd/systemd
6380 * First release with the journal
6381 http://0pointer.de/blog/projects/the-journal.html
6383 * The journal replaces both systemd-kmsg-syslogd and
6384 systemd-stdout-bridge.
6386 * New sd_pid_get_unit() API call in libsystemd-logind
6388 * Many systemadm clean-ups
6390 * Introduce remote-fs-pre.target which is ordered before all
6391 remote mounts and may be used to start services before all
6394 * Added Mageia support
6396 * Add bash completion for systemd-loginctl
6398 * Actively monitor PID file creation for daemons which exit in
6399 the parent process before having finished writing the PID
6400 file in the daemon process. Daemons which do this need to be
6401 fixed (i.e. PID file creation must have finished before the
6402 parent exits), but we now react a bit more gracefully to them.
6404 * Add colourful boot output, mimicking the well-known output
6405 of existing distributions.
6407 * New option PassCredentials= for socket units, for
6408 compatibility with a recent kernel ABI breakage.
6410 * /etc/rc.local is now hooked in via a generator binary, and
6411 thus will no longer act as synchronization point during
6414 * systemctl list-unit-files now supports --root=.
6416 * systemd-tmpfiles now understands two new commands: z, Z for
6417 relabelling files according to the SELinux database. This is
6418 useful to apply SELinux labels to specific files in /sys,
6421 * Output of SysV services is now forwarded to both the console
6422 and the journal by default, not only just the console.
6424 * New man pages for all APIs from libsystemd-login.
6426 * The build tree got reorganized and a the build system is a
6427 lot more modular allowing embedded setups to specifically
6428 select the components of systemd they are interested in.
6430 * Support for Linux systems lacking the kernel VT subsystem is
6433 * configure's --with-rootdir= got renamed to
6434 --with-rootprefix= to follow the naming used by udev and
6437 * Unless specified otherwise we will now install to /usr instead
6438 of /usr/local by default.
6440 * Processes with '@' in argv[0][0] are now excluded from the
6441 final shut-down killing spree, following the logic explained
6443 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
6445 * All processes remaining in a service cgroup when we enter
6446 the START or START_PRE states are now killed with
6447 SIGKILL. That means it is no longer possible to spawn
6448 background processes from ExecStart= lines (which was never
6449 supported anyway, and bad style).
6451 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
6452 reloading of units together.
6454 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
6455 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
6456 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
6457 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
6458 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek