1 systemd System and Service Manager
5 * A new "systemd-timesyncd" daemon has been added for
6 snchronizing the system clock across the network. It
7 implements an SNTP client. In contrast to NTP
8 implementations such as chrony or the NTP reference server
9 this only implements a client side, and does not bother with
10 the full NTP complexity, focussing only on querying time
11 from one remote server and synchronizing the local clock to
12 it. Unless you intend to serve NTP to networked clients or
13 want to connect to local hardware clocks this simple NTP
14 client should be more than appropriate for most
15 installations. The daemon runs with minimal privileges, and
16 has been hooked up with networkd to only operate when
17 network connectivity is available. The daemon saves the
18 current clock to disk every time a new NTP sync has been
19 acquired, and uses this to possibly correct the system clock
20 early at bootup, in order to accomodate for systems that
21 lack an RTC such as the Raspberry Pi and embedded devices,
22 and make sure that time monotonically progresses on these
23 systems, even if it is not always correct.
25 * The "seqnum" interface of libudev has been disabled, as it
26 was generally incompatible with device namspacing as
27 sequence numbers of devices go "missing" if the devices are
28 part of a different namespace.
30 * "systemctl list-timers" and "systemctl list-sockets" gained
31 a --recursive switch for showing units of these types also
32 for all local containers, similar in style to the
33 --recursive switch "systemctl list-units" already supports.
35 * A new RebootArgument= setting has been added for service
36 units, which may be used to specify a kernel reboot argument
37 for usage when triggering reboots with StartLimitAction=.
39 * A new FailureAction= setting has been added for service
40 units which may be used to specify an operation to trigger
41 when a service fails. This works similar to
42 StartLimitAction=, but unlike it controls what is done
43 immediately rather than only after several attempts to
44 restart the service in question.
46 * hostnamed got updated to also expose the kernel name,
47 release and version on the bus. This is useful for executing
48 commands like hostnamectl with the -H switch.
50 * The bootchart tool can now show cgroup information in the
53 * The CFS CPU quota cgroup attribute is now exposed for
54 services. The new CPUQuota= switch has been added for this
55 which takes a percentage value. Setting this will have the
56 result that a service may never get more CPU time than the
57 specified percentage, even if the machine is otherwise idle.
59 * systemd-networkd learned IPIP and SIT tunnel support.
61 * LSB init scripts exposing a dependency on $network will now
62 get a dependency on network-online.target rather than simply
63 network.target. This should bring LSB handling closer to
64 what it was on SysV systems.
66 * A new fsck.repair= kernel option has been added to control
67 how fsck shall deal with unclean file systems at boot.
69 * The (.ini) configuration file parser will now silently
70 ignore sections whose name begins with "X-". This may be
71 used to maintain application-specific extension sections in unit
74 * machined gained a new API to query the IP addresses of
75 registered containers. "machinectl status" has been updated
76 to show these addresses in its output.
78 * A new call sd_uid_get_display() has been added to the
79 sd-login APIs for querying the "primary" session of a
80 user. The "primary" session of the user is elected from the
81 user's sessions and generally a graphical session is
82 preferred over a text one.
84 * A minimal systemd-resolved daemon has been added. It
85 currently simply acts as a companion to systemd-networkd and
86 manages resolv.conf based on per-interface DNS
87 configuration, possibly supplied via DHCP. In the long run
88 we hope to extend this into a local DNSSEC enabled DNS and
91 * Two new service options StartupCPUShares= and
92 StartupBlockIOWeight= have been added that work similar to
93 CPUShares= and BlockIOWeight= however only apply during
94 system startup. This is useful to priorize certain services
95 differently during bootup than during normal runtime.
97 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
98 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
99 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
100 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
101 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
102 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
103 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
104 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
105 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
106 Marcel Holtmann, Matthew Monaco, Michael Marineau, Michael
107 Olbrich, Michal Sekletar, Mike Gilbert, Nis Martensen, Patrik
108 Flykt, Philip Lorenz, poma, Ray Strode, Robert Milasan, Scott
109 Thrasher, Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas
110 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut
111 Tezduyar Lindskog, WaLyong Cho, Will Woods, Zbigniew
114 -- Beijing, 2014-05-24
118 * When restoring the screen brightness at boot, stay away from
119 the darkest setting or from the lowest 5% of the available
120 range, depending on which is the larger value of both. This
121 should effectively protect the user from rebooting into a
122 black screen, should the brightness have been set to minimum
125 * sd-login gained a new sd_machine_get_class() call to
126 determine the class ("vm" or "container") of a machine
127 registered with machined.
129 * sd-login gained new calls
130 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
131 to query the identity of the peer of a local AF_UNIX
132 connection. They operate similar to their sd_pid_get_xyz()
135 * PID 1 will now maintain a system-wide system state engine
136 with the states "starting", "running", "degraded",
137 "maintenance", "stopping". These states are bound to system
138 startup, normal runtime, runtime with at least one failed
139 service, rescue/emergency mode and system shutdown. This
140 state is shown in the "systemctl status" output when no unit
141 name is passed. It is useful to determine system state, in
142 particularly when doing so for many systems or containers at
145 * A new command "list-machines" has been added to "systemctl"
146 that lists all local OS containers and shows their system
147 state (see above), if systemd runs inside of them.
149 * systemctl gained a new "-r" switch to recursively enumerate
150 units on all local containers, when used with the
151 "list-unit" command (which is the default one that is
152 executed when no parameters are specified).
154 * The GPT automatic partition discovery logic will now honour
155 two GPT partition flags: one may be set on a partition to
156 cause it to be mounted read-only, and the other may be set
157 on a partition to ignore it during automatic discovery.
159 * Two new GPT type UUIDs have been added for automatic root
160 partition discovery, for 32-bit and 64-bit ARM. This is not
161 particularly useful for discovering the root directory on
162 these architectures during bare-metal boots (since UEFI is
163 not common there), but still very useful to allow booting of
164 ARM disk images in nspawn with the -i option.
166 * MAC addresses of interfaces created with nspawn's
167 --network-interface= switch will now be generated from the
168 machine name, and thus be stable between multiple invocations
171 * logind will now automatically remove all IPC objects owned
172 by a user if she or he fully logs out. This makes sure that
173 users who are logged out cannot continue to consume IPC
174 resources. This covers SysV memory, semaphores and message
175 queues as well as POSIX shared memory and message
176 queues. Traditionally, SysV and POSIX IPC had no life-cycle
177 limits. With this functionality, that is corrected. This may
178 be turned off by using the RemoveIPC= switch of logind.conf.
180 * The systemd-machine-id-setup and tmpfiles tools gained a
181 --root= switch to operate on a specific root directory,
184 * journald can now forward logged messages to the TTYs of all
185 logged in users ("wall"). This is the default for all
186 emergency messages now.
188 * A new tool systemd-journal-remote has been added to stream
189 journal log messages across the network.
191 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
192 controller trees are mounted into it. Note that the
193 directories mounted beneath it are not read-only. This is a
194 security measure and is particularly useful because glibc
195 actually includes a search logic to pick any tmpfs it can
196 find to implement shm_open() if /dev/shm is not available
197 (which it might very well be in namespaced setups).
199 * machinectl gained a new "poweroff" command to cleanly power
200 down a local OS container.
202 * The PrivateDevices= unit file setting will now also drop the
203 CAP_MKNOD capability from the capability bound set, and
204 imply DevicePolicy=closed.
206 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
207 comprehensively on all long-running systemd services where
210 * systemd-udevd will now run in a disassociated mount
211 namespace. To mount directories from udev rules, make sure to
212 pull in mount units via SYSTEMD_WANTS properties.
214 * The kdbus support gained support for uploading policy into
215 the kernel. sd-bus gained support for creating "monitoring"
216 connections that can eavesdrop into all bus communication
217 for debugging purposes.
219 * Timestamps may now be specified in seconds since the UNIX
220 epoch Jan 1st, 1970 by specifying "@" followed by the value
223 * Native tcpwrap support in systemd has been removed. tcpwrap
224 is old code, not really maintained anymore and has serious
225 shortcomings, and better options such as firewalls
226 exist. For setups that require tcpwrap usage, please
227 consider invoking your socket-activated service via tcpd,
228 like on traditional inetd.
230 * A new system.conf configuration option
231 DefaultTimerAccuracySec= has been added that controls the
232 default AccuracySec= setting of .timer units.
234 * Timer units gained a new WakeSystem= switch. If enabled,
235 timers configured this way will cause the system to resume
236 from system suspend (if the system supports that, which most
239 * Timer units gained a new Persistent= switch. If enabled,
240 timers configured this way will save to disk when they have
241 been last triggered. This information is then used on next
242 reboot to possible execute overdue timer events, that
243 could not take place because the system was powered off.
244 This enables simple anacron-like behaviour for timer units.
246 * systemctl's "list-timers" will now also list the time a
247 timer unit was last triggered in addition to the next time
248 it will be triggered.
250 * systemd-networkd will now assign predictable IPv4LL
251 addresses to its local interfaces.
253 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
254 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
255 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
256 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
257 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
258 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
259 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
260 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
263 -- Berlin, 2014-03-25
267 * A new unit file setting RestrictAddressFamilies= has been
268 added to restrict which socket address families unit
269 processes gain access to. This takes address family names
270 like "AF_INET" or "AF_UNIX", and is useful to minimize the
271 attack surface of services via exotic protocol stacks. This
272 is built on seccomp system call filters.
274 * Two new unit file settings RuntimeDirectory= and
275 RuntimeDirectoryMode= have been added that may be used to
276 manage a per-daemon runtime directories below /run. This is
277 an alternative for setting up directory permissions with
278 tmpfiles snippets, and has the advantage that the runtime
279 directory's lifetime is bound to the daemon runtime and that
280 the daemon starts up with an empty directory each time. This
281 is particularly useful when writing services that drop
282 privileges using the User= or Group= setting.
284 * The DeviceAllow= unit setting now supports globbing for
285 matching against device group names.
287 * The systemd configuration file system.conf gained new
288 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
289 DefaultMemoryAccounting= to globally turn on/off accounting
290 for specific resources (cgroups) for all units. These
291 settings may still be overridden individually in each unit
294 * systemd-gpt-auto-generator is now able to discover /srv and
295 root partitions in addition to /home and swap partitions. It
296 also supports LUKS-encrypted partitions now. With this in
297 place, automatic discovery of partitions to mount following
298 the Discoverable Partitions Specification
299 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
300 is now a lot more complete. This allows booting without
301 /etc/fstab and without root= on the kernel command line on
302 systems prepared appropriately.
304 * systemd-nspawn gained a new --image= switch which allows
305 booting up disk images and Linux installations on any block
306 device that follow the Discoverable Partitions Specification
307 (see above). This means that installations made with
308 appropriately updated installers may now be started and
309 deployed using container managers, completely
310 unmodified. (We hope that libvirt-lxc will add support for
311 this feature soon, too.)
313 * systemd-nspawn gained a new --network-macvlan= setting to
314 set up a private macvlan interface for the
315 container. Similar, systemd-networkd gained a new
316 Kind=macvlan setting in .netdev files.
318 * systemd-networkd now supports configuring local addresses
321 * A new tool systemd-network-wait-online has been added to
322 synchronously wait for network connectivity using
325 * The sd-bus.h bus API gained a new sd_bus_track object for
326 tracking the life-cycle of bus peers. Note that sd-bus.h is
327 still not a public API though (unless you specify
328 --enable-kdbus on the configure command line, which however
329 voids your warranty and you get no API stability guarantee).
331 * The $XDG_RUNTIME_DIR runtime directories for each user are
332 now individual tmpfs instances, which has the benefit of
333 introducing separate pools for each user, with individual
334 size limits, and thus making sure that unprivileged clients
335 can no longer negatively impact the system or other users by
336 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
337 RuntimeDirectorySize= has been introduced that allows
338 controlling the default size limit for all users. It
339 defaults to 10% of the available physical memory. This is no
340 replacement for quotas on tmpfs though (which the kernel
341 still does not support), as /dev/shm and /tmp are still
342 shared resources used by both the system and unprivileged
345 * logind will now automatically turn off automatic suspending
346 on laptop lid close when more than one display is
347 connected. This was previously expected to be implemented
348 individually in desktop environments (such as GNOME),
349 however has been added to logind now, in order to fix a
350 boot-time race where a desktop environment might not have
351 been started yet and thus not been able to take an inhibitor
352 lock at the time where logind already suspends the system
355 * logind will now wait at least 30s after each system
356 suspend/resume cycle, and 3min after system boot before
357 suspending the system due to a closed laptop lid. This
358 should give USB docking stations and similar enough time to
359 be probed and configured after system resume and boot in
360 order to then act as suspend blocker.
362 * systemd-run gained a new --property= setting which allows
363 initialization of resource control properties (and others)
364 for the created scope or service unit. Example: "systemd-run
365 --property=BlockIOWeight=10 updatedb" may be used to run
366 updatedb at a low block IO scheduling weight.
368 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
369 now also work in --scope mode.
371 * When systemd is compiled with kdbus support, basic support
372 for enforced policies is now in place. (Note that enabling
373 kdbus still voids your warranty and no API compatibility
376 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
377 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
378 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
379 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
380 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
381 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
382 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
383 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
384 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
385 Zbigniew Jędrzejewski-Szmek
387 -- Berlin, 2014-03-12
391 * systemd will now relabel /dev after loading the SMACK policy
392 according to SMACK rules.
394 * A new unit file option AppArmorProfile= has been added to
395 set the AppArmor profile for the processes of a unit.
397 * A new condition check ConditionArchitecture= has been added
398 to conditionalize units based on the system architecture, as
399 reported by uname()'s "machine" field.
401 * systemd-networkd now supports matching on the system
402 virtualization, architecture, kernel command line, host name
405 * logind is now a lot more aggressive when suspending the
406 machine due to a closed laptop lid. Instead of acting only
407 on the lid close action, it will continuously watch the lid
408 status and act on it. This is useful for laptops where the
409 power button is on the outside of the chassis so that it can
410 be reached without opening the lid (such as the Lenovo
411 Yoga). On those machines, logind will now immediately
412 re-suspend the machine if the power button has been
413 accidentally pressed while the laptop was suspended and in a
416 * logind will now watch SW_DOCK switches and inhibit reaction
417 to the lid switch if it is pressed. This means that logind
418 will not suspend the machine anymore if the lid is closed
419 and the system is docked, if the laptop supports SW_DOCK
420 notifications via the input layer. Note that ACPI docking
421 stations do not generate this currently. Also note that this
422 logic is usually not fully sufficient and Desktop
423 Environments should take a lid switch inhibitor lock when an
424 external display is connected, as systemd will not watch
427 * nspawn will now make use of the devices cgroup controller by
428 default, and only permit creation of and access to the usual
429 API device nodes like /dev/null or /dev/random, as well as
430 access to (but not creation of) the pty devices.
432 * We will now ship a default .network file for
433 systemd-networkd that automatically configures DHCP for
434 network interfaces created by nspawn's --network-veth or
435 --network-bridge= switches.
437 * systemd will now understand the usual M, K, G, T suffixes
438 according to SI conventions (i.e. to the base 1000) when
439 referring to throughput and hardware metrics. It will stay
440 with IEC conventions (i.e. to the base 1024) for software
441 metrics, according to what is customary according to
442 Wikipedia. We explicitly document which base applies for
443 each configuration option.
445 * The DeviceAllow= setting in unit files now supports a syntax
446 to whitelist an entire group of devices node majors at once,
447 based on the /proc/devices listing. For example, with the
448 string "char-pts", it is now possible to whitelist all
449 current and future pseudo-TTYs at once.
451 * sd-event learned a new "post" event source. Event sources of
452 this type are triggered by the dispatching of any event
453 source of a type that is not "post". This is useful for
454 implementing clean-up and check event sources that are
455 triggered by other work being done in the program.
457 * systemd-networkd is no longer statically enabled, but uses
458 the usual [Install] sections so that it can be
459 enabled/disabled using systemctl. It still is enabled by
462 * When creating a veth interface pair with systemd-nspawn, the
463 host side will now be prefixed with "vb-" if
464 --network-bridge= is used, and with "ve-" if --network-veth
465 is used. This way, it is easy to distinguish these cases on
466 the host, for example to apply different configuration to
467 them with systemd-networkd.
469 * The compatibility libraries for libsystemd-journal.so,
470 libsystem-id128.so, libsystemd-login.so and
471 libsystemd-daemon.so do not make use of IFUNC
472 anymore. Instead, we now build libsystemd.so multiple times
473 under these alternative names. This means that the footprint
474 is drastically increased, but given that these are
475 transitional compatibility libraries, this should not matter
476 much. This change has been made necessary to support the ARM
477 platform for these compatibility libraries, as the ARM
478 toolchain is not really at the same level as the toolchain
479 for other architectures like x86 and does not support
480 IFUNC. Please make sure to use --enable-compat-libs only
481 during a transitional period!
483 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
484 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
485 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
486 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
487 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
488 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
489 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
490 Zbigniew Jędrzejewski-Szmek
492 -- Berlin, 2014-02-24
496 * A new component "systemd-networkd" has been added that can
497 be used to configure local network interfaces statically or
498 via DHCP. It is capable of bringing up bridges, VLANs, and
499 bonding. Currently, no hook-ups for interactive network
500 configuration are provided. Use this for your initrd,
501 container, embedded, or server setup if you need a simple,
502 yet powerful, network configuration solution. This
503 configuration subsystem is quite nifty, as it allows wildcard
504 hotplug matching in interfaces. For example, with a single
505 configuration snippet, you can configure that all Ethernet
506 interfaces showing up are automatically added to a bridge,
507 or similar. It supports link-sensing and more.
509 * A new tool "systemd-socket-proxyd" has been added which can
510 act as a bidirectional proxy for TCP sockets. This is
511 useful for adding socket activation support to services that
512 do not actually support socket activation, including virtual
513 machines and the like.
515 * Add a new tool to save/restore rfkill state on
518 * Save/restore state of keyboard backlights in addition to
519 display backlights on shutdown/boot.
521 * udev learned a new SECLABEL{} construct to label device
522 nodes with a specific security label when they appear. For
523 now, only SECLABEL{selinux} is supported, but the syntax is
524 prepared for additional security frameworks.
526 * udev gained a new scheme to configure link-level attributes
527 from files in /etc/systemd/network/*.link. These files can
528 match against MAC address, device path, driver name and type,
529 and will apply attributes like the naming policy, link speed,
530 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
531 address assignment policy (randomized, ...).
533 * The configuration of network interface naming rules for
534 "permanent interface names" has changed: a new NamePolicy=
535 setting in the [Link] section of .link files determines the
536 priority of possible naming schemes (onboard, slot, mac,
537 path). The default value of this setting is determined by
538 /usr/lib/net/links/99-default.link. Old
539 80-net-name-slot.rules udev configuration file has been
540 removed, so local configuration overriding this file should
541 be adapated to override 99-default.link instead.
543 * When the User= switch is used in a unit file, also
544 initialize $SHELL= based on the user database entry.
546 * systemd no longer depends on libdbus. All communication is
547 now done with sd-bus, systemd's low-level bus library
550 * kdbus support has been added to PID 1 itself. When kdbus is
551 enabled, this causes PID 1 to set up the system bus and
552 enable support for a new ".busname" unit type that
553 encapsulates bus name activation on kdbus. It works a little
554 bit like ".socket" units, except for bus names. A new
555 generator has been added that converts classic dbus1 service
556 activation files automatically into native systemd .busname
559 * sd-bus: add a light-weight vtable implementation that allows
560 defining objects on the bus with a simple static const
561 vtable array of its methods, signals and properties.
563 * systemd will not generate or install static dbus
564 introspection data anymore to /usr/share/dbus-1/interfaces,
565 as the precise format of these files is unclear, and
566 nothing makes use of it.
568 * A proxy daemon is now provided to proxy clients connecting
569 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
570 compatibility with classic D-Bus.
572 * A bus driver implementation has been added that supports the
573 classic D-Bus bus driver calls on kdbus, also for
574 compatibility purposes.
576 * A new API "sd-event.h" has been added that implements a
577 minimal event loop API built around epoll. It provides a
578 couple of features that direct epoll usage is lacking:
579 prioritization of events, scales to large numbers of timer
580 events, per-event timer slack (accuracy), system-wide
581 coalescing of timer events, exit handlers, watchdog
582 supervision support using systemd's sd_notify() API, child
585 * A new API "sd-rntl.h" has been added that provides an API
586 around the route netlink interface of the kernel, similar in
589 * A new API "sd-dhcp-client.h" has been added that provides a
590 small DHCPv4 client-side implementation. This is used by
593 * There is a new kernel command line option
594 "systemd.restore_state=0|1". When set to "0", none of the
595 systemd tools will restore saved runtime state to hardware
596 devices. More specifically, the rfkill and backlight states
599 * The FsckPassNo= compatibility option in mount/service units
600 has been removed. The fstab generator will now add the
601 necessary dependencies automatically, and does not require
602 PID1's support for that anymore.
604 * journalctl gained a new switch, --list-boots, that lists
605 recent boots with their times and boot IDs.
607 * The various tools like systemctl, loginctl, timedatectl,
608 busctl, systemd-run, ... have gained a new switch "-M" to
609 connect to a specific, local OS container (as direct
610 connection, without requiring SSH). This works on any
611 container that is registered with machined, such as those
612 created by libvirt-lxc or nspawn.
614 * systemd-run and systemd-analyze also gained support for "-H"
615 to connect to remote hosts via SSH. This is particularly
616 useful for systemd-run because it enables queuing of jobs
619 * machinectl gained a new command "login" to open a getty
620 login in any local container. This works with any container
621 that is registered with machined (such as those created by
622 libvirt-lxc or nspawn), and which runs systemd inside.
624 * machinectl gained a new "reboot" command that may be used to
625 trigger a reboot on a specific container that is registered
626 with machined. This works on any container that runs an init
629 * systemctl gained a new "list-timers" command to print a nice
630 listing of installed timer units with the times they elapse
633 * Alternative reboot() parameters may now be specified on the
634 "systemctl reboot" command line and are passed to the
635 reboot() system call.
637 * systemctl gained a new --job-mode= switch to configure the
638 mode to queue a job with. This is a more generic version of
639 --fail, --irreversible, and --ignore-dependencies, which are
640 still available but not advertised anymore.
642 * /etc/systemd/system.conf gained new settings to configure
643 various default timeouts of units, as well as the default
644 start limit interval and burst. These may still be overridden
647 * PID1 will now export on the bus profile data of the security
648 policy upload process (such as the SELinux policy upload to
651 * journald: when forwarding logs to the console, include
652 timestamps (following the setting in
653 /sys/module/printk/parameters/time).
655 * OnCalendar= in timer units now understands the special
656 strings "yearly" and "annually". (Both are equivalent)
658 * The accuracy of timer units is now configurable with the new
659 AccuracySec= setting. It defaults to 1min.
661 * A new dependency type JoinsNamespaceOf= has been added that
662 allows running two services within the same /tmp and network
663 namespace, if PrivateNetwork= or PrivateTmp= are used.
665 * A new command "cat" has been added to systemctl. It outputs
666 the original unit file of a unit, and concatenates the
667 contents of additional "drop-in" unit file snippets, so that
668 the full configuration is shown.
670 * systemctl now supports globbing on the various "list-xyz"
671 commands, like "list-units" or "list-sockets", as well as on
672 those commands which take multiple unit names.
674 * journalctl's --unit= switch gained support for globbing.
676 * All systemd daemons now make use of the watchdog logic so
677 that systemd automatically notices when they hang.
679 * If the $container_ttys environment variable is set,
680 getty-generator will automatically spawn a getty for each
681 listed tty. This is useful for container managers to request
682 login gettys to be spawned on as many ttys as needed.
684 * %h, %s, %U specifier support is not available anymore when
685 used in unit files for PID 1. This is because NSS calls are
686 not safe from PID 1. They stay available for --user
687 instances of systemd, and as special case for the root user.
689 * loginctl gained a new "--no-legend" switch to turn off output
692 * The "sd-login.h" API gained three new calls:
693 sd_session_is_remote(), sd_session_get_remote_user(),
694 sd_session_get_remote_host() to query information about
697 * The udev hardware database now also carries vendor/product
698 information of SDIO devices.
700 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
701 determine whether watchdog notifications are requested by
704 * Socket-activated per-connection services now include a
705 short description of the connection parameters in the
708 * tmpfiles gained a new "--boot" option. When this is not used,
709 only lines where the command character is not suffixed with
710 "!" are executed. When this option is specified, those
711 options are executed too. This partitions tmpfiles
712 directives into those that can be safely executed at any
713 time, and those which should be run only at boot (for
714 example, a line that creates /run/nologin).
716 * A new API "sd-resolve.h" has been added which provides a simple
717 asynchronous wrapper around glibc NSS host name resolution
718 calls, such as getaddrinfo(). In contrast to glibc's
719 getaddrinfo_a(), it does not use signals. In contrast to most
720 other asynchronous name resolution libraries, this one does
721 not reimplement DNS, but reuses NSS, so that alternate
722 host name resolution systems continue to work, such as mDNS,
723 LDAP, etc. This API is based on libasyncns, but it has been
724 cleaned up for inclusion in systemd.
726 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
727 "sd-daemon.h" are no longer found in individual libraries
728 libsystemd-journal.so, libsystemd-login.so,
729 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
730 merged them into a single library, libsystemd.so, which
731 provides all symbols. The reason for this is cyclic
732 dependencies, as these libraries tend to use each other's
733 symbols. So far, we have managed to workaround that by linking
734 a copy of a good part of our code into each of these
735 libraries again and again, which, however, makes certain
736 things hard to do, like sharing static variables. Also, it
737 substantially increases footprint. With this change, there
738 is only one library for the basic APIs systemd
739 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
740 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
741 library as well, however are subject to the --enable-kdbus
742 switch (see below). Note that "sd-dhcp-client.h" is not part
743 of this library (this is because it only consumes, never
744 provides, services of/to other APIs). To make the transition
745 easy from the separate libraries to the unified one, we
746 provide the --enable-compat-libs compile-time switch which
747 will generate stub libraries that are compatible with the
748 old ones but redirect all calls to the new one.
750 * All of the kdbus logic and the new APIs "sd-bus.h",
751 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
752 and "sd-utf8.h" are compile-time optional via the
753 "--enable-kdbus" switch, and they are not compiled in by
754 default. To make use of kdbus, you have to explicitly enable
755 the switch. Note however, that neither the kernel nor the
756 userspace API for all of this is considered stable yet. We
757 want to maintain the freedom to still change the APIs for
758 now. By specifying this build-time switch, you acknowledge
759 that you are aware of the instability of the current
762 * Also, note that while kdbus is pretty much complete,
763 it lacks one thing: proper policy support. This means you
764 can build a fully working system with all features; however,
765 it will be highly insecure. Policy support will be added in
766 one of the next releases, at the same time that we will
767 declare the APIs stable.
769 * When the kernel command-line argument "kdbus" is specified,
770 systemd will automatically load the kdbus.ko kernel module. At
771 this stage of development, it is only useful for testing kdbus
772 and should not be used in production. Note: if "--enable-kdbus"
773 is specified, and the kdbus.ko kernel module is available, and
774 "kdbus" is added to the kernel command line, the entire system
775 runs with kdbus instead of dbus-daemon, with the above mentioned
776 problem of missing the system policy enforcement. Also a future
777 version of kdbus.ko or a newer systemd will not be compatible with
778 each other, and will unlikely be able to boot the machine if only
779 one of them is updated.
781 * systemctl gained a new "import-environment" command which
782 uploads the caller's environment (or parts thereof) into the
783 service manager so that it is inherited by services started
784 by the manager. This is useful to upload variables like
785 $DISPLAY into the user service manager.
787 * A new PrivateDevices= switch has been added to service units
788 which allows running a service with a namespaced /dev
789 directory that does not contain any device nodes for
790 physical devices. More specifically, it only includes devices
791 such as /dev/null, /dev/urandom, and /dev/zero which are API
794 * logind has been extended to support behaviour like VT
795 switching on seats that do not support a VT. This makes
796 multi-session available on seats that are not the first seat
797 (seat0), and on systems where kernel support for VTs has
798 been disabled at compile-time.
800 * If a process holds a delay lock for system sleep or shutdown
801 and fails to release it in time, we will now log its
802 identity. This makes it easier to identify processes that
803 cause slow suspends or power-offs.
805 * When parsing /etc/crypttab, support for a new key-slot=
806 option as supported by Debian is added. It allows indicating
807 which LUKS slot to use on disk, speeding up key loading.
809 * The sd_journald_sendv() API call has been checked and
810 officially declared to be async-signal-safe so that it may
811 be invoked from signal handlers for logging purposes.
813 * Boot-time status output is now enabled automatically after a
814 short timeout if boot does not progress, in order to give
815 the user an indication what she or he is waiting for.
817 * The boot-time output has been improved to show how much time
818 remains until jobs expire.
820 * The KillMode= switch in service units gained a new possible
821 value "mixed". If set, and the unit is shut down, then the
822 initial SIGTERM signal is sent only to the main daemon
823 process, while the following SIGKILL signal is sent to
824 all remaining processes of the service.
826 * When a scope unit is registered, a new property "Controller"
827 may be set. If set to a valid bus name, systemd will send a
828 RequestStop() signal to this name when it would like to shut
829 down the scope. This may be used to hook manager logic into
830 the shutdown logic of scope units. Also, scope units may now
831 be put in a special "abandoned" state, in which case the
832 manager process which created them takes no further
833 responsibilities for it.
835 * When reading unit files, systemd will now verify
836 the access mode of these files, and warn about certain
837 suspicious combinations. This has been added to make it
838 easier to track down packaging bugs where unit files are
839 marked executable or world-writable.
841 * systemd-nspawn gained a new "--setenv=" switch to set
842 container-wide environment variables. The similar option in
843 systemd-activate was renamed from "--environment=" to
844 "--setenv=" for consistency.
846 * systemd-nspawn has been updated to create a new kdbus domain
847 for each container that is invoked, thus allowing each
848 container to have its own set of system and user buses,
849 independent of the host.
851 * systemd-nspawn gained a new --drop-capability= switch to run
852 the container with less capabilities than the default. Both
853 --drop-capability= and --capability= now take the special
854 string "all" for dropping or keeping all capabilities.
856 * systemd-nspawn gained new switches for executing containers
857 with specific SELinux labels set.
859 * systemd-nspawn gained a new --quiet switch to not generate
860 any additional output but the container's own console
863 * systemd-nspawn gained a new --share-system switch to run a
864 container without PID namespacing enabled.
866 * systemd-nspawn gained a new --register= switch to control
867 whether the container is registered with systemd-machined or
868 not. This is useful for containers that do not run full
869 OS images, but only specific apps.
871 * systemd-nspawn gained a new --keep-unit which may be used
872 when invoked as the only program from a service unit, and
873 results in registration of the unit service itself in
874 systemd-machined, instead of a newly opened scope unit.
876 * systemd-nspawn gained a new --network-interface= switch for
877 moving arbitrary interfaces to the container. The new
878 --network-veth switch creates a virtual Ethernet connection
879 between host and container. The new --network-bridge=
880 switch then allows assigning the host side of this virtual
881 Ethernet connection to a bridge device.
883 * systemd-nspawn gained a new --personality= switch for
884 setting the kernel personality for the container. This is
885 useful when running a 32-bit container on a 64-bit host. A
886 similar option Personality= is now also available for service
889 * logind will now also track a "Desktop" identifier for each
890 session which encodes the desktop environment of it. This is
891 useful for desktop environments that want to identify
892 multiple running sessions of itself easily.
894 * A new SELinuxContext= setting for service units has been
895 added that allows setting a specific SELinux execution
896 context for a service.
898 * Most systemd client tools will now honour $SYSTEMD_LESS for
899 settings of the "less" pager. By default, these tools will
900 override $LESS to allow certain operations to work, such as
901 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
902 influence this logic.
904 * systemd's "seccomp" hook-up has been changed to make use of
905 the libseccomp library instead of using its own
906 implementation. This has benefits for portability among
909 * For usage together with SystemCallFilter=, a new
910 SystemCallErrorNumber= setting has been introduced that
911 allows configuration of a system error number to be returned
912 on filtered system calls, instead of immediately killing the
913 process. Also, SystemCallArchitectures= has been added to
914 limit access to system calls of a particular architecture
915 (in order to turn off support for unused secondary
916 architectures). There is also a global
917 SystemCallArchitectures= setting in system.conf now to turn
918 off support for non-native system calls system-wide.
920 * systemd requires a kernel with a working name_to_handle_at(),
921 please see the kernel config requirements in the README file.
923 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
924 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
925 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
926 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
927 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
928 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
929 Elia Pinto, Florian Weimer, George McCollister, Goffredo
930 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
931 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
932 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
933 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
934 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
935 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
936 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
937 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
938 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
939 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
940 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
941 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
942 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
943 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
944 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
945 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
946 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
948 -- Berlin, 2014-02-20
952 * logind has gained support for facilitating privileged input
953 and drm device access for unprivileged clients. This work is
954 useful to allow Wayland display servers (and similar
955 programs, such as kmscon) to run under the user's ID and
956 access input and drm devices which are normally
957 protected. When this is used (and the kernel is new enough)
958 logind will "mute" IO on the file descriptors passed to
959 Wayland as long as it is in the background and "unmute" it
960 if it returns into the foreground. This allows secure
961 session switching without allowing background sessions to
962 eavesdrop on input and display data. This also introduces
963 session switching support if VT support is turned off in the
964 kernel, and on seats that are not seat0.
966 * A new kernel command line option luks.options= is understood
967 now which allows specifiying LUKS options for usage for LUKS
968 encrypted partitions specified with luks.uuid=.
970 * tmpfiles.d(5) snippets may now use specifier expansion in
971 path names. More specifically %m, %b, %H, %v, are now
972 replaced by the local machine id, boot id, hostname, and
973 kernel version number.
975 * A new tmpfiles.d(5) command "m" has been introduced which
976 may be used to change the owner/group/access mode of a file
977 or directory if it exists, but do nothing if it does not.
979 * This release removes high-level support for the
980 MemorySoftLimit= cgroup setting. The underlying kernel
981 cgroup attribute memory.soft_limit= is currently badly
982 designed and likely to be removed from the kernel API in its
983 current form, hence we should not expose it for now.
985 * The memory.use_hierarchy cgroup attribute is now enabled for
986 all cgroups systemd creates in the memory cgroup
987 hierarchy. This option is likely to be come the built-in
988 default in the kernel anyway, and the non-hierarchial mode
989 never made much sense in the intrinsically hierarchial
992 * A new field _SYSTEMD_SLICE= is logged along with all journal
993 messages containing the slice a message was generated
994 from. This is useful to allow easy per-customer filtering of
995 logs among other things.
997 * systemd-journald will no longer adjust the group of journal
998 files it creates to the "systemd-journal" group. Instead we
999 rely on the journal directory to be owned by the
1000 "systemd-journal" group, and its setgid bit set, so that the
1001 kernel file system layer will automatically enforce that
1002 journal files inherit this group assignment. The reason for
1003 this change is that we cannot allow NSS look-ups from
1004 journald which would be necessary to resolve
1005 "systemd-journal" to a numeric GID, because this might
1006 create deadlocks if NSS involves synchronous queries to
1007 other daemons (such as nscd, or sssd) which in turn are
1008 logging clients of journald and might block on it, which
1009 would then dead lock. A tmpfiles.d(5) snippet included in
1010 systemd will make sure the setgid bit and group are
1011 properly set on the journal directory if it exists on every
1012 boot. However, we recommend adjusting it manually after
1013 upgrades too (or from RPM scriptlets), so that the change is
1014 not delayed until next reboot.
1016 * Backlight and random seed files in /var/lib/ have moved into
1017 the /var/lib/systemd/ directory, in order to centralize all
1018 systemd generated files in one directory.
1020 * Boot time performance measurements (as displayed by
1021 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
1022 performance information if that's available to determine how
1023 much time BIOS and boot loader initialization required. With
1024 a sufficiently new BIOS you hence no longer need to boot
1025 with Gummiboot to get access to such information.
1027 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
1028 Cristian Rodríguez, Dave Reisner, David Herrmann, David
1029 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
1030 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
1031 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
1032 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
1033 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1035 -- Berlin, 2013-10-02
1039 * The Restart= option for services now understands a new
1040 on-watchdog setting, which will restart the service
1041 automatically if the service stops sending out watchdog keep
1042 alive messages (as configured with WatchdogSec=).
1044 * The getty generator (which is responsible for bringing up a
1045 getty on configured serial consoles) will no longer only
1046 start a getty on the primary kernel console but on all
1047 others, too. This makes the order in which console= is
1048 specified on the kernel command line less important.
1050 * libsystemd-logind gained a new sd_session_get_vt() call to
1051 retrieve the VT number of a session.
1053 * If the option "tries=0" is set for an entry of /etc/crypttab
1054 its passphrase is queried indefinitely instead of any
1055 maximum number of tries.
1057 * If a service with a configure PID file terminates its PID
1058 file will now be removed automatically if it still exists
1059 afterwards. This should put an end to stale PID files.
1061 * systemd-run will now also take relative binary path names
1062 for execution and no longer insists on absolute paths.
1064 * InaccessibleDirectories= and ReadOnlyDirectories= now take
1065 paths that are optionally prefixed with "-" to indicate that
1066 it should not be considered a failure if they do not exist.
1068 * journalctl -o (and similar commands) now understands a new
1069 output mode "short-precise", it is similar to "short" but
1070 shows timestamps with usec accuracy.
1072 * The option "discard" (as known from Debian) is now
1073 synonymous to "allow-discards" in /etc/crypttab. In fact,
1074 "discard" is preferred now (since it is easier to remember
1077 * Some licensing clean-ups were made, so that more code is now
1078 LGPL-2.1 licensed than before.
1080 * A minimal tool to save/restore the display backlight
1081 brightness across reboots has been added. It will store the
1082 backlight setting as late as possible at shutdown, and
1083 restore it as early as possible during reboot.
1085 * A logic to automatically discover and enable home and swap
1086 partitions on GPT disks has been added. With this in place
1087 /etc/fstab becomes optional for many setups as systemd can
1088 discover certain partitions located on the root disk
1089 automatically. Home partitions are recognized under their
1090 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
1091 partitions are recognized under their GPT type ID
1092 0657fd6da4ab43c484e50933c84b4f4f.
1094 * systemd will no longer pass any environment from the kernel
1095 or initrd to system services. If you want to set an
1096 environment for all services, do so via the kernel command
1097 line systemd.setenv= assignment.
1099 * The systemd-sysctl tool no longer natively reads the file
1100 /etc/sysctl.conf. If desired, the file should be symlinked
1101 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
1102 legacy support by a symlink rather than built-in code, it
1103 also makes the otherwise hidden order of application of the
1104 different files visible. (Note that this partly reverts to a
1105 pre-198 application order of sysctl knobs!)
1107 * The "systemctl set-log-level" and "systemctl dump" commands
1108 have been moved to systemd-analyze.
1110 * systemd-run learned the new --remain-after-exit switch,
1111 which causes the scope unit not to be cleaned up
1112 automatically after the process terminated.
1114 * tmpfiles learned a new --exclude-prefix= switch to exclude
1115 certain paths from operation.
1117 * journald will now automatically flush all messages to disk
1118 as soon as a message of the log priorities CRIT, ALERT or
1121 Contributions from: Andrew Cook, Brandon Philips, Christian
1122 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
1123 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
1124 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
1125 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
1126 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
1127 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
1128 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
1129 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
1130 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
1131 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
1132 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
1133 William Giokas, Zbigniew Jędrzejewski-Szmek
1135 -- Berlin, 2013-09-13
1139 * The documentation has been updated to cover the various new
1140 concepts introduced with 205.
1142 * Unit files now understand the new %v specifier which
1143 resolves to the kernel version string as returned by "uname
1146 * systemctl now supports filtering the unit list output by
1147 load state, active state and sub state, using the new
1150 * "systemctl status" will now show the results of the
1151 condition checks (like ConditionPathExists= and similar) of
1152 the last start attempts of the unit. They are also logged to
1155 * "journalctl -b" may now be used to look for boot output of a
1156 specific boot. Try "journalctl -b -1" for the previous boot,
1157 but the syntax is substantially more powerful.
1159 * "journalctl --show-cursor" has been added which prints the
1160 cursor string the last shown log line. This may then be used
1161 with the new "journalctl --after-cursor=" switch to continue
1162 browsing logs from that point on.
1164 * "journalctl --force" may now be used to force regeneration
1167 * Creation of "dead" device nodes has been moved from udev
1168 into kmod and tmpfiles. Previously, udev would read the kmod
1169 databases to pre-generate dead device nodes based on meta
1170 information contained in kernel modules, so that these would
1171 be auto-loaded on access rather then at boot. As this
1172 does not really have much to do with the exposing actual
1173 kernel devices to userspace this has always been slightly
1174 alien in the udev codebase. Following the new scheme kmod
1175 will now generate a runtime snippet for tmpfiles from the
1176 module meta information and it now is tmpfiles' job to the
1177 create the nodes. This also allows overriding access and
1178 other parameters for the nodes using the usual tmpfiles
1179 facilities. As side effect this allows us to remove the
1180 CAP_SYS_MKNOD capability bit from udevd entirely.
1182 * logind's device ACLs may now be applied to these "dead"
1183 devices nodes too, thus finally allowing managed access to
1184 devices such as /dev/snd/sequencer whithout loading the
1185 backing module right-away.
1187 * A new RPM macro has been added that may be used to apply
1188 tmpfiles configuration during package installation.
1190 * systemd-detect-virt and ConditionVirtualization= now can
1191 detect User-Mode-Linux machines (UML).
1193 * journald will now implicitly log the effective capabilities
1194 set of processes in the message metadata.
1196 * systemd-cryptsetup has gained support for TrueCrypt volumes.
1198 * The initrd interface has been simplified (more specifically,
1199 support for passing performance data via environment
1200 variables and fsck results via files in /run has been
1201 removed). These features were non-essential, and are
1202 nowadays available in a much nicer way by having systemd in
1203 the initrd serialize its state and have the hosts systemd
1204 deserialize it again.
1206 * The udev "keymap" data files and tools to apply keyboard
1207 specific mappings of scan to key codes, and force-release
1208 scan code lists have been entirely replaced by a udev
1209 "keyboard" builtin and a hwdb data file.
1211 * systemd will now honour the kernel's "quiet" command line
1212 argument also during late shutdown, resulting in a
1213 completely silent shutdown when used.
1215 * There's now an option to control the SO_REUSEPORT socket
1216 option in .socket units.
1218 * Instance units will now automatically get a per-template
1219 subslice of system.slice unless something else is explicitly
1220 configured. For example, instances of sshd@.service will now
1221 implicitly be placed in system-sshd.slice rather than
1222 system.slice as before.
1224 * Test coverage support may now be enabled at build time.
1226 Contributions from: Dave Reisner, Frederic Crozat, Harald
1227 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
1228 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
1229 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
1230 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
1231 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
1232 Giokas, Zbigniew Jędrzejewski-Szmek
1234 -- Berlin, 2013-07-23
1238 * Two new unit types have been introduced:
1240 Scope units are very similar to service units, however, are
1241 created out of pre-existing processes -- instead of PID 1
1242 forking off the processes. By using scope units it is
1243 possible for system services and applications to group their
1244 own child processes (worker processes) in a powerful way
1245 which then maybe used to organize them, or kill them
1246 together, or apply resource limits on them.
1248 Slice units may be used to partition system resources in an
1249 hierarchial fashion and then assign other units to them. By
1250 default there are now three slices: system.slice (for all
1251 system services), user.slice (for all user sessions),
1252 machine.slice (for VMs and containers).
1254 Slices and scopes have been introduced primarily in
1255 context of the work to move cgroup handling to a
1256 single-writer scheme, where only PID 1
1257 creates/removes/manages cgroups.
1259 * There's a new concept of "transient" units. In contrast to
1260 normal units these units are created via an API at runtime,
1261 not from configuration from disk. More specifically this
1262 means it is now possible to run arbitrary programs as
1263 independent services, with all execution parameters passed
1264 in via bus APIs rather than read from disk. Transient units
1265 make systemd substantially more dynamic then it ever was,
1266 and useful as a general batch manager.
1268 * logind has been updated to make use of scope and slice units
1269 for managing user sessions. As a user logs in he will get
1270 his own private slice unit, to which all sessions are added
1271 as scope units. We also added support for automatically
1272 adding an instance of user@.service for the user into the
1273 slice. Effectively logind will no longer create cgroup
1274 hierarchies on its own now, it will defer entirely to PID 1
1275 for this by means of scope, service and slice units. Since
1276 user sessions this way become entities managed by PID 1
1277 the output of "systemctl" is now a lot more comprehensive.
1279 * A new mini-daemon "systemd-machined" has been added which
1280 may be used by virtualization managers to register local
1281 VMs/containers. nspawn has been updated accordingly, and
1282 libvirt will be updated shortly. machined will collect a bit
1283 of meta information about the VMs/containers, and assign
1284 them their own scope unit (see above). The collected
1285 meta-data is then made available via the "machinectl" tool,
1286 and exposed in "ps" and similar tools. machined/machinectl
1287 is compile-time optional.
1289 * As discussed earlier, the low-level cgroup configuration
1290 options ControlGroup=, ControlGroupModify=,
1291 ControlGroupPersistent=, ControlGroupAttribute= have been
1292 removed. Please use high-level attribute settings instead as
1293 well as slice units.
1295 * A new bus call SetUnitProperties() has been added to alter
1296 various runtime parameters of a unit. This is primarily
1297 useful to alter cgroup parameters dynamically in a nice way,
1298 but will be extended later on to make more properties
1299 modifiable at runtime. systemctl gained a new set-properties
1300 command that wraps this call.
1302 * A new tool "systemd-run" has been added which can be used to
1303 run arbitrary command lines as transient services or scopes,
1304 while configuring a number of settings via the command
1305 line. This tool is currently very basic, however already
1306 very useful. We plan to extend this tool to even allow
1307 queuing of execution jobs with time triggers from the
1308 command line, similar in fashion to "at".
1310 * nspawn will now inform the user explicitly that kernels with
1311 audit enabled break containers, and suggest the user to turn
1314 * Support for detecting the IMA and AppArmor security
1315 frameworks with ConditionSecurity= has been added.
1317 * journalctl gained a new "-k" switch for showing only kernel
1318 messages, mimicking dmesg output; in addition to "--user"
1319 and "--system" switches for showing only user's own logs
1322 * systemd-delta can now show information about drop-in
1323 snippets extending unit files.
1325 * libsystemd-bus has been substantially updated but is still
1326 not available as public API.
1328 * systemd will now look for the "debug" argument on the kernel
1329 command line and enable debug logging, similar to
1330 "systemd.log_level=debug" already did before.
1332 * "systemctl set-default", "systemctl get-default" has been
1333 added to configure the default.target symlink, which
1334 controls what to boot into by default.
1336 * "systemctl set-log-level" has been added as a convenient
1337 way to raise and lower systemd logging threshold.
1339 * "systemd-analyze plot" will now show the time the various
1340 generators needed for execution, as well as information
1341 about the unit file loading.
1343 * libsystemd-journal gained a new sd_journal_open_files() call
1344 for opening specific journal files. journactl also gained a
1345 new switch to expose this new functionality. Previously we
1346 only supported opening all files from a directory, or all
1347 files from the system, as opening individual files only is
1348 racy due to journal file rotation.
1350 * systemd gained the new DefaultEnvironment= setting in
1351 /etc/systemd/system.conf to set environment variables for
1354 * If a privileged process logs a journal message with the
1355 OBJECT_PID= field set, then journald will automatically
1356 augment this with additional OBJECT_UID=, OBJECT_GID=,
1357 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
1358 system services want to log events about specific client
1359 processes. journactl/systemctl has been updated to make use
1360 of this information if all log messages regarding a specific
1363 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
1364 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
1365 Reisner, David Coppa, David King, David Strauss, Eelco
1366 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
1367 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
1368 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
1369 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
1370 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
1371 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
1372 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
1373 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
1374 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
1375 Łukasz Stelmach, 장동준
1379 * The Python bindings gained some minimal support for the APIs
1380 exposed by libsystemd-logind.
1382 * ConditionSecurity= gained support for detecting SMACK. Since
1383 this condition already supports SELinux and AppArmor we only
1384 miss IMA for this. Patches welcome!
1386 Contributions from: Karol Lewandowski, Lennart Poettering,
1387 Zbigniew Jędrzejewski-Szmek
1391 * systemd-nspawn will now create /etc/resolv.conf if
1392 necessary, before bind-mounting the host's file onto it.
1394 * systemd-nspawn will now store meta information about a
1395 container on the container's cgroup as extended attribute
1396 fields, including the root directory.
1398 * The cgroup hierarchy has been reworked in many ways. All
1399 objects any of the components systemd creates in the cgroup
1400 tree are now suffixed. More specifically, user sessions are
1401 now placed in cgroups suffixed with ".session", users in
1402 cgroups suffixed with ".user", and nspawn containers in
1403 cgroups suffixed with ".nspawn". Furthermore, all cgroup
1404 names are now escaped in a simple scheme to avoid collision
1405 of userspace object names with kernel filenames. This work
1406 is preparation for making these objects relocatable in the
1407 cgroup tree, in order to allow easy resource partitioning of
1408 these objects without causing naming conflicts.
1410 * systemctl list-dependencies gained the new switches
1411 --plain, --reverse, --after and --before.
1413 * systemd-inhibit now shows the process name of processes that
1414 have taken an inhibitor lock.
1416 * nss-myhostname will now also resolve "localhost"
1417 implicitly. This makes /etc/hosts an optional file and
1418 nicely handles that on IPv6 ::1 maps to both "localhost" and
1421 * libsystemd-logind.so gained a new call
1422 sd_get_machine_names() to enumerate running containers and
1423 VMs (currently only supported by very new libvirt and
1424 nspawn). sd_login_monitor can now be used to watch
1425 VMs/containers coming and going.
1427 * .include is not allowed recursively anymore, and only in
1428 unit files. Usually it is better to use drop-in snippets in
1429 .d/*.conf anyway, as introduced with systemd 198.
1431 * systemd-analyze gained a new "critical-chain" command that
1432 determines the slowest chain of units run during system
1433 boot-up. It is very useful for tracking down where
1434 optimizing boot time is the most beneficial.
1436 * systemd will no longer allow manipulating service paths in
1437 the name=systemd:/system cgroup tree using ControlGroup= in
1438 units. (But is still fine with it in all other dirs.)
1440 * There's a new systemd-nspawn@.service service file that may
1441 be used to easily run nspawn containers as system
1442 services. With the container's root directory in
1443 /var/lib/container/foobar it is now sufficient to run
1444 "systemctl start systemd-nspawn@foobar.service" to boot it.
1446 * systemd-cgls gained a new parameter "--machine" to list only
1447 the processes within a certain container.
1449 * ConditionSecurity= now can check for "apparmor". We still
1450 are lacking checks for SMACK and IMA for this condition
1451 check though. Patches welcome!
1453 * A new configuration file /etc/systemd/sleep.conf has been
1454 added that may be used to configure which kernel operation
1455 systemd is supposed to execute when "suspend", "hibernate"
1456 or "hybrid-sleep" is requested. This makes the new kernel
1457 "freeze" state accessible to the user.
1459 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
1460 the passed argument if applicable.
1462 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
1463 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
1464 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
1465 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
1466 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
1467 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
1468 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
1473 * The output of 'systemctl list-jobs' got some polishing. The
1474 '--type=' argument may now be passed more than once. A new
1475 command 'systemctl list-sockets' has been added which shows
1476 a list of kernel sockets systemd is listening on with the
1477 socket units they belong to, plus the units these socket
1480 * The experimental libsystemd-bus library got substantial
1481 updates to work in conjunction with the (also experimental)
1482 kdbus kernel project. It works well enough to exchange
1483 messages with some sophistication. Note that kdbus is not
1484 ready yet, and the library is mostly an elaborate test case
1485 for now, and not installable.
1487 * systemd gained a new unit 'systemd-static-nodes.service'
1488 that generates static device nodes earlier during boot, and
1489 can run in conjunction with udev.
1491 * libsystemd-login gained a new call sd_pid_get_user_unit()
1492 to retrieve the user systemd unit a process is running
1493 in. This is useful for systems where systemd is used as
1496 * systemd-nspawn now places all containers in the new /machine
1497 top-level cgroup directory in the name=systemd
1498 hierarchy. libvirt will soon do the same, so that we get a
1499 uniform separation of /system, /user and /machine for system
1500 services, user processes and containers/virtual
1501 machines. This new cgroup hierarchy is also useful to stick
1502 stable names to specific container instances, which can be
1503 recognized later this way (this name may be controlled
1504 via systemd-nspawn's new -M switch). libsystemd-login also
1505 gained a new call sd_pid_get_machine_name() to retrieve the
1506 name of the container/VM a specific process belongs to.
1508 * bootchart can now store its data in the journal.
1510 * libsystemd-journal gained a new call
1511 sd_journal_add_conjunction() for AND expressions to the
1512 matching logic. This can be used to express more complex
1513 logical expressions.
1515 * journactl can now take multiple --unit= and --user-unit=
1518 * The cryptsetup logic now understands the "luks.key=" kernel
1519 command line switch for specifying a file to read the
1520 decryption key from. Also, if a configured key file is not
1521 found the tool will now automatically fall back to prompting
1524 * Python systemd.journal module was updated to wrap recently
1525 added functions from libsystemd-journal. The interface was
1526 changed to bring the low level interface in s.j._Reader
1527 closer to the C API, and the high level interface in
1528 s.j.Reader was updated to wrap and convert all data about
1531 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
1532 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
1533 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
1534 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
1535 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
1536 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1540 * journalctl --update-catalog now understands a new --root=
1541 option to operate on catalogs found in a different root
1544 * During shutdown after systemd has terminated all running
1545 services a final killing loop kills all remaining left-over
1546 processes. We will now print the name of these processes
1547 when we send SIGKILL to them, since this usually indicates a
1550 * If /etc/crypttab refers to password files stored on
1551 configured mount points automatic dependencies will now be
1552 generated to ensure the specific mount is established first
1553 before the key file is attempted to be read.
1555 * 'systemctl status' will now show information about the
1556 network sockets a socket unit is listening on.
1558 * 'systemctl status' will also shown information about any
1559 drop-in configuration file for units. (Drop-In configuration
1560 files in this context are files such as
1561 /etc/systemd/systemd/foobar.service.d/*.conf)
1563 * systemd-cgtop now optionally shows summed up CPU times of
1564 cgroups. Press '%' while running cgtop to switch between
1565 percentage and absolute mode. This is useful to determine
1566 which cgroups use up the most CPU time over the entire
1567 runtime of the system. systemd-cgtop has also been updated
1568 to be 'pipeable' for processing with further shell tools.
1570 * 'hostnamectl set-hostname' will now allow setting of FQDN
1573 * The formatting and parsing of time span values has been
1574 changed. The parser now understands fractional expressions
1575 such as "5.5h". The formatter will now output fractional
1576 expressions for all time spans under 1min, i.e. "5.123456s"
1577 rather than "5s 123ms 456us". For time spans under 1s
1578 millisecond values are shown, for those under 1ms
1579 microsecond values are shown. This should greatly improve
1580 all time-related output of systemd.
1582 * libsystemd-login and libsystemd-journal gained new
1583 functions for querying the poll() events mask and poll()
1584 timeout value for integration into arbitrary event
1587 * localectl gained the ability to list available X11 keymaps
1588 (models, layouts, variants, options).
1590 * 'systemd-analyze dot' gained the ability to filter for
1591 specific units via shell-style globs, to create smaller,
1592 more useful graphs. I.e. it is now possible to create simple
1593 graphs of all the dependencies between only target units, or
1594 of all units that Avahi has dependencies with.
1596 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
1597 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
1598 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
1599 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
1600 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
1601 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
1602 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
1606 * The boot-time readahead implementation for rotating media
1607 will now read the read-ahead data in multiple passes which
1608 consist of all read requests made in equidistant time
1609 intervals. This means instead of strictly reading read-ahead
1610 data in its physical order on disk we now try to find a
1611 middle ground between physical and access time order.
1613 * /etc/os-release files gained a new BUILD_ID= field for usage
1614 on operating systems that provide continuous builds of OS
1617 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
1618 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
1619 William Douglas, Zbigniew Jędrzejewski-Szmek
1623 * systemd-python gained an API exposing libsystemd-daemon.
1625 * The SMACK setup logic gained support for uploading CIPSO
1628 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
1629 ReadOnlyDirectories= and InaccessibleDirectories= has
1630 changed. The private /tmp and /var/tmp directories are now
1631 shared by all processes of a service (which means
1632 ExecStartPre= may now leave data in /tmp that ExecStart= of
1633 the same service can still access). When a service is
1634 stopped its temporary directories are immediately deleted
1635 (normal clean-up with tmpfiles is still done in addition to
1638 * By default, systemd will now set a couple of sysctl
1639 variables in the kernel: the safe sysrq options are turned
1640 on, IP route verification is turned on, and source routing
1641 disabled. The recently added hardlink and softlink
1642 protection of the kernel is turned on. These settings should
1643 be reasonably safe, and good defaults for all new systems.
1645 * The predictable network naming logic may now be turned off
1646 with a new kernel command line switch: net.ifnames=0.
1648 * A new libsystemd-bus module has been added that implements a
1649 pretty complete D-Bus client library. For details see:
1651 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
1653 * journald will now explicitly flush the journal files to disk
1654 at the latest 5min after each write. The file will then also
1655 be marked offline until the next write. This should increase
1656 reliability in case of a crash. The synchronization delay
1657 can be configured via SyncIntervalSec= in journald.conf.
1659 * There's a new remote-fs-setup.target unit that can be used
1660 to pull in specific services when at least one remote file
1661 system is to be mounted.
1663 * There are new targets timers.target and paths.target as
1664 canonical targets to pull user timer and path units in
1665 from. This complements sockets.target with a similar
1666 purpose for socket units.
1668 * libudev gained a new call udev_device_set_attribute_value()
1669 to set sysfs attributes of a device.
1671 * The udev daemon now sets the default number of worker
1672 processes executed in parallel based on the number of available
1673 CPUs instead of the amount of available RAM. This is supposed
1674 to provide a more reliable default and limit a too aggressive
1675 paralellism for setups with 1000s of devices connected.
1677 Contributions from: Auke Kok, Colin Walters, Cristian
1678 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
1679 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
1680 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
1681 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
1682 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
1683 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
1684 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
1685 Zbigniew Jędrzejewski-Szmek
1689 * Configuration of unit files may now be extended via drop-in
1690 files without having to edit/override the unit files
1691 themselves. More specifically, if the administrator wants to
1692 change one value for a service file foobar.service he can
1693 now do so by dropping in a configuration snippet into
1694 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
1695 will load all these snippets and apply them on top of the
1696 main unit configuration file, possibly extending or
1697 overriding its settings. Using these drop-in snippets is
1698 generally nicer than the two earlier options for changing
1699 unit files locally: copying the files from
1700 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
1701 them there; or creating a new file in /etc/systemd/system/
1702 that incorporates the original one via ".include". Drop-in
1703 snippets into these .d/ directories can be placed in any
1704 directory systemd looks for units in, and the usual
1705 overriding semantics between /usr/lib, /etc and /run apply
1708 * Most unit file settings which take lists of items can now be
1709 reset by assigning the empty string to them. For example,
1710 normally, settings such as Environment=FOO=BAR append a new
1711 environment variable assignment to the environment block,
1712 each time they are used. By assigning Environment= the empty
1713 string the environment block can be reset to empty. This is
1714 particularly useful with the .d/*.conf drop-in snippets
1715 mentioned above, since this adds the ability to reset list
1716 settings from vendor unit files via these drop-ins.
1718 * systemctl gained a new "list-dependencies" command for
1719 listing the dependencies of a unit recursively.
1721 * Inhibitors are now honored and listed by "systemctl
1722 suspend", "systemctl poweroff" (and similar) too, not only
1723 GNOME. These commands will also list active sessions by
1726 * Resource limits (as exposed by the various control group
1727 controllers) can now be controlled dynamically at runtime
1728 for all units. More specifically, you can now use a command
1729 like "systemctl set-cgroup-attr foobar.service cpu.shares
1730 2000" to alter the CPU shares a specific service gets. These
1731 settings are stored persistently on disk, and thus allow the
1732 administrator to easily adjust the resource usage of
1733 services with a few simple commands. This dynamic resource
1734 management logic is also available to other programs via the
1735 bus. Almost any kernel cgroup attribute and controller is
1738 * systemd-vconsole-setup will now copy all font settings to
1739 all allocated VTs, where it previously applied them only to
1742 * libsystemd-login gained the new sd_session_get_tty() API
1745 * This release drops support for a few legacy or
1746 distribution-specific LSB facility names when parsing init
1747 scripts: $x-display-manager, $mail-transfer-agent,
1748 $mail-transport-agent, $mail-transfer-agent, $smtp,
1749 $null. Also, the mail-transfer-agent.target unit backing
1750 this has been removed. Distributions which want to retain
1751 compatibility with this should carry the burden for
1752 supporting this themselves and patch support for these back
1753 in, if they really need to. Also, the facilities $syslog and
1754 $local_fs are now ignored, since systemd does not support
1755 early-boot LSB init scripts anymore, and these facilities
1756 are implied anyway for normal services. syslog.target has
1759 * There are new bus calls on PID1's Manager object for
1760 cancelling jobs, and removing snapshot units. Previously,
1761 both calls were only available on the Job and Snapshot
1764 * systemd-journal-gatewayd gained SSL support.
1766 * The various "environment" files, such as /etc/locale.conf
1767 now support continuation lines with a backslash ("\") as
1768 last character in the line, similar in style (but different)
1769 to how this is supported in shells.
1771 * For normal user processes the _SYSTEMD_USER_UNIT= field is
1772 now implicitly appended to every log entry logged. systemctl
1773 has been updated to filter by this field when operating on a
1774 user systemd instance.
1776 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
1777 CAP_AUDIT_CONTROL capabilities to the capabilities set for
1778 the container. This makes it easier to boot unmodified
1779 Fedora systems in a container, which however still requires
1780 audit=0 to be passed on the kernel command line. Auditing in
1781 kernel and userspace is unfortunately still too broken in
1782 context of containers, hence we recommend compiling it out
1783 of the kernel or using audit=0. Hopefully this will be fixed
1784 one day for good in the kernel.
1786 * nspawn gained the new --bind= and --bind-ro= parameters to
1787 bind mount specific directories from the host into the
1790 * nspawn will now mount its own devpts file system instance
1791 into the container, in order not to leak pty devices from
1792 the host into the container.
1794 * systemd will now read the firmware boot time performance
1795 information from the EFI variables, if the used boot loader
1796 supports this, and takes it into account for boot performance
1797 analysis via "systemd-analyze". This is currently supported
1798 only in conjunction with Gummiboot, but could be supported
1799 by other boot loaders too. For details see:
1801 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
1803 * A new generator has been added that automatically mounts the
1804 EFI System Partition (ESP) to /boot, if that directory
1805 exists, is empty, and no other file system has been
1806 configured to be mounted there.
1808 * logind will now send out PrepareForSleep(false) out
1809 unconditionally, after coming back from suspend. This may be
1810 used by applications as asynchronous notification for
1811 system resume events.
1813 * "systemctl unlock-sessions" has been added, that allows
1814 unlocking the screens of all user sessions at once, similar
1815 how "systemctl lock-sessions" already locked all users
1816 sessions. This is backed by a new D-Bus call UnlockSessions().
1818 * "loginctl seat-status" will now show the master device of a
1819 seat. (i.e. the device of a seat that needs to be around for
1820 the seat to be considered available, usually the graphics
1823 * tmpfiles gained a new "X" line type, that allows
1824 configuration of files and directories (with wildcards) that
1825 shall be excluded from automatic cleanup ("aging").
1827 * udev default rules set the device node permissions now only
1828 at "add" events, and do not change them any longer with a
1829 later "change" event.
1831 * The log messages for lid events and power/sleep keypresses
1832 now carry a message ID.
1834 * We now have a substantially larger unit test suite, but this
1835 continues to be work in progress.
1837 * udevadm hwdb gained a new --root= parameter to change the
1838 root directory to operate relative to.
1840 * logind will now issue a background sync() request to the kernel
1841 early at shutdown, so that dirty buffers are flushed to disk early
1842 instead of at the last moment, in order to optimize shutdown
1845 * A new bootctl tool has been added that is an interface for
1846 certain boot loader operations. This is currently a preview
1847 and is likely to be extended into a small mechanism daemon
1848 like timedated, localed, hostnamed, and can be used by
1849 graphical UIs to enumerate available boot options, and
1850 request boot into firmware operations.
1852 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
1853 the rest of the package. It also has been updated to work
1854 correctly in initrds.
1856 * Policykit previously has been runtime optional, and is now
1857 also compile time optional via a configure switch.
1859 * systemd-analyze has been reimplemented in C. Also "systemctl
1860 dot" has moved into systemd-analyze.
1862 * "systemctl status" with no further parameters will now print
1863 the status of all active or failed units.
1865 * Operations such as "systemctl start" can now be executed
1866 with a new mode "--irreversible" which may be used to queue
1867 operations that cannot accidentally be reversed by a later
1868 job queuing. This is by default used to make shutdown
1869 requests more robust.
1871 * The Python API of systemd now gained a new module for
1872 reading journal files.
1874 * A new tool kernel-install has been added that can install
1875 kernel images according to the Boot Loader Specification:
1877 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
1879 * Boot time console output has been improved to provide
1880 animated boot time output for hanging jobs.
1882 * A new tool systemd-activate has been added which can be used
1883 to test socket activation with, directly from the command
1884 line. This should make it much easier to test and debug
1885 socket activation in daemons.
1887 * journalctl gained a new "--reverse" (or -r) option to show
1888 journal output in reverse order (i.e. newest line first).
1890 * journalctl gained a new "--pager-end" (or -e) option to jump
1891 to immediately jump to the end of the journal in the
1892 pager. This is only supported in conjunction with "less".
1894 * journalctl gained a new "--user-unit=" option, that works
1895 similar to "--unit=" but filters for user units rather than
1898 * A number of unit files to ease adoption of systemd in
1899 initrds has been added. This moves some minimal logic from
1900 the various initrd implementations into systemd proper.
1902 * The journal files are now owned by a new group
1903 "systemd-journal", which exists specifically to allow access
1904 to the journal, and nothing else. Previously, we used the
1905 "adm" group for that, which however possibly covers more
1906 than just journal/log file access. This new group is now
1907 already used by systemd-journal-gatewayd to ensure this
1908 daemon gets access to the journal files and as little else
1909 as possible. Note that "make install" will also set FS ACLs
1910 up for /var/log/journal to give "adm" and "wheel" read
1911 access to it, in addition to "systemd-journal" which owns
1912 the journal files. We recommend that packaging scripts also
1913 add read access to "adm" + "wheel" to /var/log/journal, and
1914 all existing/future journal files. To normal users and
1915 administrators little changes, however packagers need to
1916 ensure to create the "systemd-journal" system group at
1917 package installation time.
1919 * The systemd-journal-gatewayd now runs as unprivileged user
1920 systemd-journal-gateway:systemd-journal-gateway. Packaging
1921 scripts need to create these system user/group at
1924 * timedated now exposes a new boolean property CanNTP that
1925 indicates whether a local NTP service is available or not.
1927 * systemd-detect-virt will now also detect xen PVs
1929 * The pstore file system is now mounted by default, if it is
1932 * In addition to the SELinux and IMA policies we will now also
1933 load SMACK policies at early boot.
1935 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
1936 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
1937 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
1938 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
1939 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
1940 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
1941 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
1942 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
1943 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
1944 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
1945 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
1946 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
1947 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
1948 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
1952 * Timer units now support calendar time events in addition to
1953 monotonic time events. That means you can now trigger a unit
1954 based on a calendar time specification such as "Thu,Fri
1955 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
1956 or fifth day of any month of the year 2013, given that it is
1957 a thursday or friday. This brings timer event support
1958 considerably closer to cron's capabilities. For details on
1959 the supported calendar time specification language see
1962 * udev now supports a number of different naming policies for
1963 network interfaces for predictable names, and a combination
1964 of these policies is now the default. Please see this wiki
1965 document for details:
1967 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
1969 * Auke Kok's bootchart implementation has been added to the
1970 systemd tree. It is an optional component that can graph the
1971 boot in quite some detail. It is one of the best bootchart
1972 implementations around and minimal in its code and
1975 * nss-myhostname has been integrated into the systemd source
1976 tree. nss-myhostname guarantees that the local hostname
1977 always stays resolvable via NSS. It has been a weak
1978 requirement of systemd-hostnamed since a long time, and
1979 since its code is actually trivial we decided to just
1980 include it in systemd's source tree. It can be turned off
1981 with a configure switch.
1983 * The read-ahead logic is now capable of properly detecting
1984 whether a btrfs file system is on SSD or rotating media, in
1985 order to optimize the read-ahead scheme. Previously, it was
1986 only capable of detecting this on traditional file systems
1989 * In udev, additional device properties are now read from the
1990 IAB in addition to the OUI database. Also, Bluetooth company
1991 identities are attached to the devices as well.
1993 * In service files %U may be used as specifier that is
1994 replaced by the configured user name of the service.
1996 * nspawn may now be invoked without a controlling TTY. This
1997 makes it suitable for invocation as its own service. This
1998 may be used to set up a simple containerized server system
1999 using only core OS tools.
2001 * systemd and nspawn can now accept socket file descriptors
2002 when they are started for socket activation. This enables
2003 implementation of socket activated nspawn
2004 containers. i.e. think about autospawning an entire OS image
2005 when the first SSH or HTTP connection is received. We expect
2006 that similar functionality will also be added to libvirt-lxc
2009 * journalctl will now suppress ANSI color codes when
2010 presenting log data.
2012 * systemctl will no longer show control group information for
2013 a unit if a the control group is empty anyway.
2015 * logind can now automatically suspend/hibernate/shutdown the
2018 * /etc/machine-info and hostnamed now also expose the chassis
2019 type of the system. This can be used to determine whether
2020 the local system is a laptop, desktop, handset or
2021 tablet. This information may either be configured by the
2022 user/vendor or is automatically determined from ACPI and DMI
2023 information if possible.
2025 * A number of PolicyKit actions are now bound together with
2026 "imply" rules. This should simplify creating UIs because
2027 many actions will now authenticate similar ones as well.
2029 * Unit files learnt a new condition ConditionACPower= which
2030 may be used to conditionalize a unit depending on whether an
2031 AC power source is connected or not, of whether the system
2032 is running on battery power.
2034 * systemctl gained a new "is-failed" verb that may be used in
2035 shell scripts and suchlike to check whether a specific unit
2036 is in the "failed" state.
2038 * The EnvironmentFile= setting in unit files now supports file
2039 globbing, and can hence be used to easily read a number of
2040 environment files at once.
2042 * systemd will no longer detect and recognize specific
2043 distributions. All distribution-specific #ifdeffery has been
2044 removed, systemd is now fully generic and
2045 distribution-agnostic. Effectively, not too much is lost as
2046 a lot of the code is still accessible via explicit configure
2047 switches. However, support for some distribution specific
2048 legacy configuration file formats has been dropped. We
2049 recommend distributions to simply adopt the configuration
2050 files everybody else uses now and convert the old
2051 configuration from packaging scripts. Most distributions
2052 already did that. If that's not possible or desirable,
2053 distributions are welcome to forward port the specific
2054 pieces of code locally from the git history.
2056 * When logging a message about a unit systemd will now always
2057 log the unit name in the message meta data.
2059 * localectl will now also discover system locale data that is
2060 not stored in locale archives, but directly unpacked.
2062 * logind will no longer unconditionally use framebuffer
2063 devices as seat masters, i.e. as devices that are required
2064 to be existing before a seat is considered preset. Instead,
2065 it will now look for all devices that are tagged as
2066 "seat-master" in udev. By default framebuffer devices will
2067 be marked as such, but depending on local systems other
2068 devices might be marked as well. This may be used to
2069 integrate graphics cards using closed source drivers (such
2070 as NVidia ones) more nicely into logind. Note however, that
2071 we recommend using the open source NVidia drivers instead,
2072 and no udev rules for the closed-source drivers will be
2073 shipped from us upstream.
2075 Contributions from: Adam Williamson, Alessandro Crismani, Auke
2076 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
2077 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
2078 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
2079 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
2080 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
2081 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
2082 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
2083 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
2084 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
2085 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
2086 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
2091 * udev gained support for loading additional device properties
2092 from an indexed database that is keyed by vendor/product IDs
2093 and similar device identifiers. For the beginning this
2094 "hwdb" is populated with data from the well-known PCI and
2095 USB database, but also includes PNP, ACPI and OID data. In
2096 the longer run this indexed database shall grow into
2097 becoming the one central database for non-essential
2098 userspace device metadata. Previously, data from the PCI/USB
2099 database was only attached to select devices, since the
2100 lookup was a relatively expensive operation due to O(n) time
2101 complexity (with n being the number of entries in the
2102 database). Since this is now O(1), we decided to add in this
2103 data for all devices where this is available, by
2104 default. Note that the indexed database needs to be rebuilt
2105 when new data files are installed. To achieve this you need
2106 to update your packaging scripts to invoke "udevadm hwdb
2107 --update" after installation of hwdb data files. For
2108 RPM-based distributions we introduced the new
2109 %udev_hwdb_update macro for this purpose.
2111 * The Journal gained support for the "Message Catalog", an
2112 indexed database to link up additional information with
2113 journal entries. For further details please check:
2115 http://www.freedesktop.org/wiki/Software/systemd/catalog
2117 The indexed message catalog database also needs to be
2118 rebuilt after installation of message catalog files. Use
2119 "journalctl --update-catalog" for this. For RPM-based
2120 distributions we introduced the %journal_catalog_update
2121 macro for this purpose.
2123 * The Python Journal bindings gained support for the standard
2124 Python logging framework.
2126 * The Journal API gained new functions for checking whether
2127 the underlying file system of a journal file is capable of
2128 properly reporting file change notifications, or whether
2129 applications that want to reflect journal changes "live"
2130 need to recheck journal files continuously in appropriate
2133 * It is now possible to set the "age" field for tmpfiles
2134 entries to 0, indicating that files matching this entry
2135 shall always be removed when the directories are cleaned up.
2137 * coredumpctl gained a new "gdb" verb which invokes gdb
2138 right-away on the selected coredump.
2140 * There's now support for "hybrid sleep" on kernels that
2141 support this, in addition to "suspend" and "hibernate". Use
2142 "systemctl hybrid-sleep" to make use of this.
2144 * logind's HandleSuspendKey= setting (and related settings)
2145 now gained support for a new "lock" setting to simply
2146 request the screen lock on all local sessions, instead of
2147 actually executing a suspend or hibernation.
2149 * systemd will now mount the EFI variables file system by
2152 * Socket units now gained support for configuration of the
2153 SMACK security label.
2155 * timedatectl will now output the time of the last and next
2156 daylight saving change.
2158 * We dropped support for various legacy and distro-specific
2159 concepts, such as insserv, early-boot SysV services
2160 (i.e. those for non-standard runlevels such as 'b' or 'S')
2161 or ArchLinux /etc/rc.conf support. We recommend the
2162 distributions who still need support this to either continue
2163 to maintain the necessary patches downstream, or find a
2164 different solution. (Talk to us if you have questions!)
2166 * Various systemd components will now bypass PolicyKit checks
2167 for root and otherwise handle properly if PolicyKit is not
2168 found to be around. This should fix most issues for
2169 PolicyKit-less systems. Quite frankly this should have been
2170 this way since day one. It is absolutely our intention to
2171 make systemd work fine on PolicyKit-less systems, and we
2172 consider it a bug if something does not work as it should if
2173 PolicyKit is not around.
2175 * For embedded systems it is now possible to build udev and
2176 systemd without blkid and/or kmod support.
2178 * "systemctl switch-root" is now capable of switching root
2179 more than once. I.e. in addition to transitions from the
2180 initrd to the host OS it is now possible to transition to
2181 further OS images from the host. This is useful to implement
2182 offline updating tools.
2184 * Various other additions have been made to the RPM macros
2185 shipped with systemd. Use %udev_rules_update() after
2186 installing new udev rules files. %_udevhwdbdir,
2187 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
2188 %_sysctldir are now available which resolve to the right
2189 directories for packages to place various data files in.
2191 * journalctl gained the new --full switch (in addition to
2192 --all, to disable ellipsation for long messages.
2194 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
2195 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
2196 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
2197 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
2198 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
2199 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
2200 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
2201 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
2202 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
2206 * journalctl gained new --since= and --until= switches to
2207 filter by time. It also now supports nice filtering for
2208 units via --unit=/-u.
2210 * Type=oneshot services may use ExecReload= and do the
2213 * The journal daemon now supports time-based rotation and
2214 vacuuming, in addition to the usual disk-space based
2217 * The journal will now index the available field values for
2218 each field name. This enables clients to show pretty drop
2219 downs of available match values when filtering. The bash
2220 completion of journalctl has been updated
2221 accordingly. journalctl gained a new switch -F to list all
2222 values a certain field takes in the journal database.
2224 * More service events are now written as structured messages
2225 to the journal, and made recognizable via message IDs.
2227 * The timedated, localed and hostnamed mini-services which
2228 previously only provided support for changing time, locale
2229 and hostname settings from graphical DEs such as GNOME now
2230 also have a minimal (but very useful) text-based client
2231 utility each. This is probably the nicest way to changing
2232 these settings from the command line now, especially since
2233 it lists available options and is fully integrated with bash
2236 * There's now a new tool "systemd-coredumpctl" to list and
2237 extract coredumps from the journal.
2239 * We now install a README each in /var/log/ and
2240 /etc/rc.d/init.d explaining where the system logs and init
2241 scripts went. This hopefully should help folks who go to
2242 that dirs and look into the otherwise now empty void and
2243 scratch their heads.
2245 * When user-services are invoked (by systemd --user) the
2246 $MANAGERPID env var is set to the PID of systemd.
2248 * SIGRTMIN+24 when sent to a --user instance will now result
2249 in immediate termination of systemd.
2251 * gatewayd received numerous feature additions such as a
2252 "follow" mode, for live syncing and filtering.
2254 * browse.html now allows filtering and showing detailed
2255 information on specific entries. Keyboard navigation and
2256 mouse screen support has been added.
2258 * gatewayd/journalctl now supports HTML5/JSON
2259 Server-Sent-Events as output.
2261 * The SysV init script compatibility logic will now
2262 heuristically determine whether a script supports the
2263 "reload" verb, and only then make this available as
2266 * "systemctl status --follow" has been removed, use "journalctl
2269 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
2270 have been removed since they are hardly useful to be
2273 * And I'd like to take the opportunity to specifically mention
2274 Zbigniew for his great contributions. Zbigniew, you rock!
2276 Contributions from: Andrew Eikum, Christian Hesse, Colin
2277 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
2278 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
2279 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
2280 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
2281 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
2282 Jędrzejewski-Szmek, Сковорода Никита Андреевич
2286 * If /etc/vconsole.conf is non-existent or empty we will no
2287 longer load any console font or key map at boot by
2288 default. Instead the kernel defaults will be left
2289 intact. This is definitely the right thing to do, as no
2290 configuration should mean no configuration, and hard-coding
2291 font names that are different on all archs is probably a bad
2292 idea. Also, the kernel default key map and font should be
2293 good enough for most cases anyway, and mostly identical to
2294 the userspace fonts/key maps we previously overloaded them
2295 with. If distributions want to continue to default to a
2296 non-kernel font or key map they should ship a default
2297 /etc/vconsole.conf with the appropriate contents.
2299 Contributions from: Colin Walters, Daniel J Walsh, Dave
2300 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
2301 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2305 * journalctl gained a new --cursor= switch to show entries
2306 starting from the specified location in the journal.
2308 * We now enforce a size limit on journal entry fields exported
2309 with "-o json" in journalctl. Fields larger than 4K will be
2310 assigned null. This can be turned off with --all.
2312 * An (optional) journal gateway daemon is now available as
2313 "systemd-journal-gatewayd.service". This service provides
2314 access to the journal via HTTP and JSON. This functionality
2315 will be used to implement live log synchronization in both
2316 pull and push modes, but has various other users too, such
2317 as easy log access for debugging of embedded devices. Right
2318 now it is already useful to retrieve the journal via HTTP:
2320 # systemctl start systemd-journal-gatewayd.service
2321 # wget http://localhost:19531/entries
2323 This will download the journal contents in a
2324 /var/log/messages compatible format. The same as JSON:
2326 # curl -H"Accept: application/json" http://localhost:19531/entries
2328 This service is also accessible via a web browser where a
2329 single static HTML5 app is served that uses the JSON logic
2330 to enable the user to do some basic browsing of the
2331 journal. This will be extended later on. Here's an example
2332 screenshot of this app in its current state:
2334 http://0pointer.de/public/journal-gatewayd
2336 Contributions from: Kay Sievers, Lennart Poettering, Robert
2337 Milasan, Tom Gundersen
2341 * The bash completion logic is now available for journalctl
2344 * We do not mount the "cpuset" controller anymore together with
2345 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
2346 started if no parameters are assigned to it. "cpuset" hence
2347 broke code that assumed it it could create "cpu" groups and
2350 * journalctl -f will now subscribe to terminal size changes,
2351 and line break accordingly.
2353 Contributions from: Dave Reisner, Kay Sievers, Lennart
2354 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
2358 * nspawn will now create a symlink /etc/localtime in the
2359 container environment, copying the host's timezone
2360 setting. Previously this has been done via a bind mount, but
2361 since symlinks cannot be bind mounted this has now been
2362 changed to create/update the appropriate symlink.
2364 * journalctl -n's line number argument is now optional, and
2365 will default to 10 if omitted.
2367 * journald will now log the maximum size the journal files may
2368 take up on disk. This is particularly useful if the default
2369 built-in logic of determining this parameter from the file
2370 system size is used. Use "systemctl status
2371 systemd-journald.service" to see this information.
2373 * The multi-seat X wrapper tool has been stripped down. As X
2374 is now capable of enumerating graphics devices via udev in a
2375 seat-aware way the wrapper is not strictly necessary
2376 anymore. A stripped down temporary stop-gap is still shipped
2377 until the upstream display managers have been updated to
2378 fully support the new X logic. Expect this wrapper to be
2379 removed entirely in one of the next releases.
2381 * HandleSleepKey= in logind.conf has been split up into
2382 HandleSuspendKey= and HandleHibernateKey=. The old setting
2383 is not available anymore. X11 and the kernel are
2384 distuingishing between these keys and we should too. This
2385 also means the inhibition lock for these keys has been split
2388 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
2389 Poettering, Lukas Nykryn, Václav Pavlín
2393 * Whenever a unit changes state we will now log this to the
2394 journal and show along the unit's own log output in
2397 * ConditionPathIsMountPoint= can now properly detect bind
2398 mount points too. (Previously, a bind mount of one file
2399 system to another place in the same file system could not be
2400 detected as mount, since they shared struct stat's st_dev
2403 * We will now mount the cgroup controllers cpu, cpuacct,
2404 cpuset and the controllers net_cls, net_prio together by
2407 * nspawn containers will now have a virtualized boot
2408 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
2409 over with a randomized ID at container initialization). This
2410 has the effect of making "journalctl -b" do the right thing
2413 * The JSON output journal serialization has been updated not
2414 to generate "endless" list objects anymore, but rather one
2415 JSON object per line. This is more in line how most JSON
2416 parsers expect JSON objects. The new output mode
2417 "json-pretty" has been added to provide similar output, but
2418 neatly aligned for readability by humans.
2420 * We dropped all explicit sync() invocations in the shutdown
2421 code. The kernel does this implicitly anyway in the kernel
2422 reboot() syscall. halt(8)'s -n option is now a compatibility
2425 * We now support virtualized reboot() in containers, as
2426 supported by newer kernels. We will fall back to exit() if
2427 CAP_SYS_REBOOT is not available to the container. Also,
2428 nspawn makes use of this now and will actually reboot the
2429 container if the containerized OS asks for that.
2431 * journalctl will only show local log output by default
2432 now. Use --merge (-m) to show remote log output, too.
2434 * libsystemd-journal gained the new sd_journal_get_usage()
2435 call to determine the current disk usage of all journal
2436 files. This is exposed in the new "journalctl --disk-usage"
2439 * journald gained a new configuration setting SplitMode= in
2440 journald.conf which may be used to control how user journals
2441 are split off. See journald.conf(5) for details.
2443 * A new condition type ConditionFileNotEmpty= has been added.
2445 * tmpfiles' "w" lines now support file globbing, to write
2446 multiple files at once.
2448 * We added Python bindings for the journal submission
2449 APIs. More Python APIs for a number of selected APIs will
2450 likely follow. Note that we intend to add native bindings
2451 only for the Python language, as we consider it common
2452 enough to deserve bindings shipped within systemd. There are
2453 various projects outside of systemd that provide bindings
2454 for languages such as PHP or Lua.
2456 * Many conditions will now resolve specifiers such as %i. In
2457 addition, PathChanged= and related directives of .path units
2458 now support specifiers as well.
2460 * There's now a new RPM macro definition for the system preset
2463 * journald will now warn if it ca not forward a message to the
2464 syslog daemon because its socket is full.
2466 * timedated will no longer write or process /etc/timezone,
2467 except on Debian. As we do not support late mounted /usr
2468 anymore /etc/localtime always being a symlink is now safe,
2469 and hence the information in /etc/timezone is not necessary
2472 * logind will now always reserve one VT for a text getty (VT6
2473 by default). Previously if more than 6 X sessions where
2474 started they took up all the VTs with auto-spawned gettys,
2475 so that no text gettys were available anymore.
2477 * udev will now automatically inform the btrfs kernel logic
2478 about btrfs RAID components showing up. This should make
2479 simple hotplug based btrfs RAID assembly work.
2481 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
2482 (but not for its children which will stay at the kernel
2483 default). This should allow setups with a lot more listening
2486 * systemd will now always pass the configured timezone to the
2487 kernel at boot. timedated will do the same when the timezone
2490 * logind's inhibition logic has been updated. By default,
2491 logind will now handle the lid switch, the power and sleep
2492 keys all the time, even in graphical sessions. If DEs want
2493 to handle these events on their own they should take the new
2494 handle-power-key, handle-sleep-key and handle-lid-switch
2495 inhibitors during their runtime. A simple way to achiveve
2496 that is to invoke the DE wrapped in an invocation of:
2498 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
2500 * Access to unit operations is now checked via SELinux taking
2501 the unit file label and client process label into account.
2503 * systemd will now notify the administrator in the journal
2504 when he over-mounts a non-empty directory.
2506 * There are new specifiers that are resolved in unit files,
2507 for the host name (%H), the machine ID (%m) and the boot ID
2510 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
2511 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
2512 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
2513 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
2514 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
2515 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2516 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
2520 * Support for reading structured kernel messages from
2521 /dev/kmsg has now been added and is enabled by default.
2523 * Support for reading kernel messages from /proc/kmsg has now
2524 been removed. If you want kernel messages in the journal
2525 make sure to run a recent kernel (>= 3.5) that supports
2526 reading structured messages from /dev/kmsg (see
2527 above). /proc/kmsg is now exclusive property of classic
2528 syslog daemons again.
2530 * The libudev API gained the new
2531 udev_device_new_from_device_id() call.
2533 * The logic for file system namespace (ReadOnlyDirectory=,
2534 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
2535 require pivot_root() anymore. This means fewer temporary
2536 directories are created below /tmp for this feature.
2538 * nspawn containers will now see and receive all submounts
2539 made on the host OS below the root file system of the
2542 * Forward Secure Sealing is now supported for Journal files,
2543 which provide cryptographical sealing of journal files so
2544 that attackers cannot alter log history anymore without this
2545 being detectable. Lennart will soon post a blog story about
2546 this explaining it in more detail.
2548 * There are two new service settings RestartPreventExitStatus=
2549 and SuccessExitStatus= which allow configuration of exit
2550 status (exit code or signal) which will be excepted from the
2551 restart logic, resp. consider successful.
2553 * journalctl gained the new --verify switch that can be used
2554 to check the integrity of the structure of journal files and
2555 (if Forward Secure Sealing is enabled) the contents of
2558 * nspawn containers will now be run with /dev/stdin, /dev/fd/
2559 and similar symlinks pre-created. This makes running shells
2560 as container init process a lot more fun.
2562 * The fstab support can now handle PARTUUID= and PARTLABEL=
2565 * A new ConditionHost= condition has been added to match
2566 against the hostname (with globs) and machine ID. This is
2567 useful for clusters where a single OS image is used to
2568 provision a large number of hosts which shall run slightly
2569 different sets of services.
2571 * Services which hit the restart limit will now be placed in a
2574 Contributions from: Bertram Poettering, Dave Reisner, Huang
2575 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
2576 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
2580 * When running in --user mode systemd will now become a
2581 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
2582 tree a lot more organized.
2584 * A new PartOf= unit dependency type has been introduced that
2585 may be used to group services in a natural way.
2587 * "systemctl enable" may now be used to enable instances of
2590 * journalctl now prints error log levels in red, and
2591 warning/notice log levels in bright white. It also supports
2592 filtering by log level now.
2594 * cgtop gained a new -n switch (similar to top), to configure
2595 the maximum number of iterations to run for. It also gained
2596 -b, to run in batch mode (accepting no input).
2598 * The suffix ".service" may now be omitted on most systemctl
2599 command lines involving service unit names.
2601 * There's a new bus call in logind to lock all sessions, as
2602 well as a loginctl verb for it "lock-sessions".
2604 * libsystemd-logind.so gained a new call sd_journal_perror()
2605 that works similar to libc perror() but logs to the journal
2606 and encodes structured information about the error number.
2608 * /etc/crypttab entries now understand the new keyfile-size=
2611 * shutdown(8) now can send a (configurable) wall message when
2612 a shutdown is cancelled.
2614 * The mount propagation mode for the root file system will now
2615 default to "shared", which is useful to make containers work
2616 nicely out-of-the-box so that they receive new mounts from
2617 the host. This can be undone locally by running "mount
2618 --make-rprivate /" if needed.
2620 * The prefdm.service file has been removed. Distributions
2621 should maintain this unit downstream if they intend to keep
2622 it around. However, we recommend writing normal unit files
2623 for display managers instead.
2625 * Since systemd is a crucial part of the OS we will now
2626 default to a number of compiler switches that improve
2627 security (hardening) such as read-only relocations, stack
2628 protection, and suchlike.
2630 * The TimeoutSec= setting for services is now split into
2631 TimeoutStartSec= and TimeoutStopSec= to allow configuration
2632 of individual time outs for the start and the stop phase of
2635 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
2636 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
2637 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
2638 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
2639 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
2640 Gundersen, Zbigniew Jędrzejewski-Szmek
2644 * The journal and id128 C APIs are now fully documented as man
2647 * Extra safety checks have been added when transitioning from
2648 the initial RAM disk to the main system to avoid accidental
2651 * /etc/crypttab entries now understand the new keyfile-offset=
2654 * systemctl -t can now be used to filter by unit load state.
2656 * The journal C API gained the new sd_journal_wait() call to
2657 make writing synchronous journal clients easier.
2659 * journalctl gained the new -D switch to show journals from a
2662 * journalctl now displays a special marker between log
2663 messages of two different boots.
2665 * The journal is now explicitly flushed to /var via a service
2666 systemd-journal-flush.service, rather than implicitly simply
2667 by seeing /var/log/journal to be writable.
2669 * journalctl (and the journal C APIs) can now match for much
2670 more complex expressions, with alternatives and
2673 * When transitioning from the initial RAM disk to the main
2674 system we will now kill all processes in a killing spree to
2675 ensure no processes stay around by accident.
2677 * Three new specifiers may be used in unit files: %u, %h, %s
2678 resolve to the user name, user home directory resp. user
2679 shell. This is useful for running systemd user instances.
2681 * We now automatically rotate journal files if their data
2682 object hash table gets a fill level > 75%. We also size the
2683 hash table based on the configured maximum file size. This
2684 together should lower hash collisions drastically and thus
2685 speed things up a bit.
2687 * journalctl gained the new "--header" switch to introspect
2688 header data of journal files.
2690 * A new setting SystemCallFilters= has been added to services
2691 which may be used to apply blacklists or whitelists to
2692 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
2694 * nspawn gained a new --link-journal= switch (and quicker: -j)
2695 to link the container journal with the host. This makes it
2696 very easy to centralize log viewing on the host for all
2697 guests while still keeping the journal files separated.
2699 * Many bugfixes and optimizations
2701 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
2702 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
2703 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
2708 * Several tools now understand kernel command line arguments,
2709 which are only read when run in an initial RAM disk. They
2710 usually follow closely their normal counterparts, but are
2713 * There's a new tool to analyze the readahead files that are
2714 automatically generated at boot. Use:
2716 /usr/lib/systemd/systemd-readahead analyze /.readahead
2718 * We now provide an early debug shell on tty9 if this enabled. Use:
2720 systemctl enable debug-shell.service
2722 * All plymouth related units have been moved into the Plymouth
2723 package. Please make sure to upgrade your Plymouth version
2726 * systemd-tmpfiles now supports getting passed the basename of
2727 a configuration file only, in which case it will look for it
2728 in all appropriate directories automatically.
2730 * udevadm info now takes a /dev or /sys path as argument, and
2731 does the right thing. Example:
2733 udevadm info /dev/sda
2734 udevadm info /sys/class/block/sda
2736 * systemctl now prints a warning if a unit is stopped but a
2737 unit that might trigger it continues to run. Example: a
2738 service is stopped but the socket that activates it is left
2741 * "systemctl status" will now mention if the log output was
2742 shortened due to rotation since a service has been started.
2744 * The journal API now exposes functions to determine the
2745 "cutoff" times due to rotation.
2747 * journald now understands SIGUSR1 and SIGUSR2 for triggering
2748 immediately flushing of runtime logs to /var if possible,
2749 resp. for triggering immediate rotation of the journal
2752 * It is now considered an error if a service is attempted to
2753 be stopped that is not loaded.
2755 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
2757 * systemd-analyze now supports Python 3
2759 * tmpfiles now supports cleaning up directories via aging
2760 where the first level dirs are always kept around but
2761 directories beneath it automatically aged. This is enabled
2762 by prefixing the age field with '~'.
2764 * Seat objects now expose CanGraphical, CanTTY properties
2765 which is required to deal with very fast bootups where the
2766 display manager might be running before the graphics drivers
2767 completed initialization.
2769 * Seat objects now expose a State property.
2771 * We now include RPM macros for service enabling/disabling
2772 based on the preset logic. We recommend RPM based
2773 distributions to make use of these macros if possible. This
2774 makes it simpler to reuse RPM spec files across
2777 * We now make sure that the collected systemd unit name is
2778 always valid when services log to the journal via
2781 * There's a new man page kernel-command-line(7) detailing all
2782 command line options we understand.
2784 * The fstab generator may now be disabled at boot by passing
2785 fstab=0 on the kernel command line.
2787 * A new kernel command line option modules-load= is now understood
2788 to load a specific kernel module statically, early at boot.
2790 * Unit names specified on the systemctl command line are now
2791 automatically escaped as needed. Also, if file system or
2792 device paths are specified they are automatically turned
2793 into the appropriate mount or device unit names. Example:
2795 systemctl status /home
2796 systemctl status /dev/sda
2798 * The SysVConsole= configuration option has been removed from
2799 system.conf parsing.
2801 * The SysV search path is no longer exported on the D-Bus
2804 * The Names= option is been removed from unit file parsing.
2806 * There's a new man page bootup(7) detailing the boot process.
2808 * Every unit and every generator we ship with systemd now
2809 comes with full documentation. The self-explanatory boot is
2812 * A couple of services gained "systemd-" prefixes in their
2813 name if they wrap systemd code, rather than only external
2814 code. Among them fsck@.service which is now
2815 systemd-fsck@.service.
2817 * The HaveWatchdog property has been removed from the D-Bus
2820 * systemd.confirm_spawn= on the kernel command line should now
2823 * There's a new man page crypttab(5) which details all options
2824 we actually understand.
2826 * systemd-nspawn gained a new --capability= switch to pass
2827 additional capabilities to the container.
2829 * timedated will now read known NTP implementation unit names
2830 from /usr/lib/systemd/ntp-units.d/*.list,
2831 systemd-timedated-ntp.target has been removed.
2833 * journalctl gained a new switch "-b" that lists log data of
2834 the current boot only.
2836 * The notify socket is in the abstract namespace again, in
2837 order to support daemons which chroot() at start-up.
2839 * There is a new Storage= configuration option for journald
2840 which allows configuration of where log data should go. This
2841 also provides a way to disable journal logging entirely, so
2842 that data collected is only forwarded to the console, the
2843 kernel log buffer or another syslog implementation.
2845 * Many bugfixes and optimizations
2847 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
2848 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
2849 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
2850 Shawn Landden, Tom Gundersen
2854 * "systemctl help <unit>" now shows the man page if one is
2857 * Several new man pages have been added.
2859 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
2860 MaxLevelConsole= can now be specified in
2861 journald.conf. These options allow reducing the amount of
2862 data stored on disk or forwarded by the log level.
2864 * TimerSlackNSec= can now be specified in system.conf for
2865 PID1. This allows system-wide power savings.
2867 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
2868 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
2873 * logind is now capable of (optionally) handling power and
2874 sleep keys as well as the lid switch.
2876 * journalctl now understands the syntax "journalctl
2877 /usr/bin/avahi-daemon" to get all log output of a specific
2880 * CapabilityBoundingSet= in system.conf now also influences
2881 the capability bound set of usermode helpers of the kernel.
2883 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
2884 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
2885 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
2886 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
2890 * Note that we skipped 139 releases here in order to set the
2891 new version to something that is greater than both udev's
2892 and systemd's most recent version number.
2894 * udev: all udev sources are merged into the systemd source tree now.
2895 All future udev development will happen in the systemd tree. It
2896 is still fully supported to use the udev daemon and tools without
2897 systemd running, like in initramfs or other init systems. Building
2898 udev though, will require the *build* of the systemd tree, but
2899 udev can be properly *run* without systemd.
2901 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
2902 should be used to create dead device nodes as workarounds for broken
2905 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
2906 no longer supported. udev_monitor_new_from_netlink() needs to be
2907 used to subscribe to events.
2909 * udev: when udevd is started by systemd, processes which are left
2910 behind by forking them off of udev rules, are unconditionally cleaned
2911 up and killed now after the event handling has finished. Services or
2912 daemons must be started as systemd services. Services can be
2913 pulled-in by udev to get started, but they can no longer be directly
2914 forked by udev rules.
2916 * udev: the daemon binary is called systemd-udevd now and installed
2917 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
2918 to adapt to that, create symlink, or rename the binary after building
2921 * libudev no longer provides these symbols:
2922 udev_monitor_from_socket()
2923 udev_queue_get_failed_list_entry()
2924 udev_get_{dev,sys,run}_path()
2925 The versions number was bumped and symbol versioning introduced.
2927 * systemd-loginctl and systemd-journalctl have been renamed
2928 to loginctl and journalctl to match systemctl.
2930 * The config files: /etc/systemd/systemd-logind.conf and
2931 /etc/systemd/systemd-journald.conf have been renamed to
2932 logind.conf and journald.conf. Package updates should rename
2933 the files to the new names on upgrade.
2935 * For almost all files the license is now LGPL2.1+, changed
2936 from the previous GPL2.0+. Exceptions are some minor stuff
2937 of udev (which will be changed to LGPL2.1 eventually, too),
2938 and the MIT licensed sd-daemon.[ch] library that is suitable
2939 to be used as drop-in files.
2941 * systemd and logind now handle system sleep states, in
2942 particular suspending and hibernating.
2944 * logind now implements a sleep/shutdown/idle inhibiting logic
2945 suitable for a variety of uses. Soonishly Lennart will blog
2946 about this in more detail.
2948 * var-run.mount and var-lock.mount are no longer provided
2949 (which prevously bind mounted these directories to their new
2950 places). Distributions which have not converted these
2951 directories to symlinks should consider stealing these files
2952 from git history and add them downstream.
2954 * We introduced the Documentation= field for units and added
2955 this to all our shipped units. This is useful to make it
2956 easier to explore the boot and the purpose of the various
2959 * All smaller setup units (such as
2960 systemd-vconsole-setup.service) now detect properly if they
2961 are run in a container and are skipped when
2962 appropriate. This guarantees an entirely noise-free boot in
2963 Linux container environments such as systemd-nspawn.
2965 * A framework for implementing offline system updates is now
2966 integrated, for details see:
2967 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
2969 * A new service type Type=idle is available now which helps us
2970 avoiding ugly interleaving of getty output and boot status
2973 * There's now a system-wide CapabilityBoundingSet= option to
2974 globally reduce the set of capabilities for the
2975 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
2976 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
2977 even CAP_NET_ADMIN system-wide for secure systems.
2979 * There are now system-wide DefaultLimitXXX= options to
2980 globally change the defaults of the various resource limits
2981 for all units started by PID 1.
2983 * Harald Hoyer's systemd test suite has been integrated into
2984 systemd which allows easy testing of systemd builds in qemu
2985 and nspawn. (This is really awesome! Ask us for details!)
2987 * The fstab parser is now implemented as generator, not inside
2990 * systemctl will now warn you if .mount units generated from
2991 /etc/fstab are out of date due to changes in fstab that
2992 have not been read by systemd yet.
2994 * systemd is now suitable for usage in initrds. Dracut has
2995 already been updated to make use of this. With this in place
2996 initrds get a slight bit faster but primarily are much
2997 easier to introspect and debug since "systemctl status" in
2998 the host system can be used to introspect initrd services,
2999 and the journal from the initrd is kept around too.
3001 * systemd-delta has been added, a tool to explore differences
3002 between user/admin configuration and vendor defaults.
3004 * PrivateTmp= now affects both /tmp and /var/tmp.
3006 * Boot time status messages are now much prettier and feature
3007 proper english language. Booting up systemd has never been
3010 * Read-ahead pack files now include the inode number of all
3011 files to pre-cache. When the inode changes the pre-caching
3012 is not attempted. This should be nicer to deal with updated
3013 packages which might result in changes of read-ahead
3016 * We now temporaritly lower the kernel's read_ahead_kb variable
3017 when collecting read-ahead data to ensure the kernel's
3018 built-in read-ahead does not add noise to our measurements
3019 of necessary blocks to pre-cache.
3021 * There's now RequiresMountsFor= to add automatic dependencies
3022 for all mounts necessary for a specific file system path.
3024 * MountAuto= and SwapAuto= have been removed from
3025 system.conf. Mounting file systems at boot has to take place
3028 * nspawn now learned a new switch --uuid= to set the machine
3029 ID on the command line.
3031 * nspawn now learned the -b switch to automatically search
3034 * vt102 is now the default TERM for serial TTYs, upgraded from
3037 * systemd-logind now works on VT-less systems.
3039 * The build tree has been reorganized. The individual
3040 components now have directories of their own.
3042 * A new condition type ConditionPathIsReadWrite= is now available.
3044 * nspawn learned the new -C switch to create cgroups for the
3045 container in other hierarchies.
3047 * We now have support for hardware watchdogs, configurable in
3050 * The scheduled shutdown logic now has a public API.
3052 * We now mount /tmp as tmpfs by default, but this can be
3053 masked and /etc/fstab can override it.
3055 * Since udisks does not make use of /media anymore we are not
3056 mounting a tmpfs on it anymore.
3058 * journalctl gained a new --local switch to only interleave
3059 locally generated journal files.
3061 * We can now load the IMA policy at boot automatically.
3063 * The GTK tools have been split off into a systemd-ui.
3065 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
3066 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
3067 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
3068 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
3069 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
3070 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
3071 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
3072 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
3073 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
3078 * This is mostly a bugfix release
3080 * Support optional initialization of the machine ID from the
3081 KVM or container configured UUID.
3083 * Support immediate reboots with "systemctl reboot -ff"
3085 * Show /etc/os-release data in systemd-analyze output
3087 * Many bugfixes for the journal, including endianness fixes and
3088 ensuring that disk space enforcement works
3090 * sd-login.h is C++ comptaible again
3092 * Extend the /etc/os-release format on request of the Debian
3095 * We now refuse non-UTF8 strings used in various configuration
3096 and unit files. This is done to ensure we do not pass invalid
3097 data over D-Bus or expose it elsewhere.
3099 * Register Mimo USB Screens as suitable for automatic seat
3102 * Read SELinux client context from journal clients in a race
3105 * Reorder configuration file lookup order. /etc now always
3106 overrides /run in order to allow the administrator to always
3107 and unconditionally override vendor supplied or
3108 automatically generated data.
3110 * The various user visible bits of the journal now have man
3111 pages. We still lack man pages for the journal API calls
3114 * We now ship all man pages in HTML format again in the
3117 Contributions from: Dave Reisner, Dirk Eibach, Frederic
3118 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
3119 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
3124 * This is mostly a bugfix release
3126 * systems lacking /etc/os-release are no longer supported.
3128 * Various functionality updates to libsystemd-login.so
3130 * Track class of PAM logins to distuingish greeters from
3133 Contributions from: Kay Sievers, Lennart Poettering, Michael
3138 * This is an important bugfix release for v41.
3140 * Building man pages is now optional which should be useful
3141 for those building systemd from git but unwilling to install
3144 * Watchdog support for supervising services is now usable. In
3145 a future release support for hardware watchdogs
3146 (i.e. /dev/watchdog) will be added building on this.
3148 * Service start rate limiting is now configurable and can be
3149 turned off per service. When a start rate limit is hit a
3150 reboot can automatically be triggered.
3152 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
3154 Contributions from: Benjamin Franzke, Bill Nottingham,
3155 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
3156 Schmidt, Michał Górny, Piotr Drąg
3160 * The systemd binary is installed /usr/lib/systemd/systemd now;
3161 An existing /sbin/init symlink needs to be adapted with the
3164 * The code that loads kernel modules has been ported to invoke
3165 libkmod directly, instead of modprobe. This means we do not
3166 support systems with module-init-tools anymore.
3168 * Watchdog support is now already useful, but still not
3171 * A new kernel command line option systemd.setenv= is
3172 understood to set system wide environment variables
3173 dynamically at boot.
3175 * We now limit the set of capabilities of systemd-journald.
3177 * We now set SIGPIPE to ignore by default, since it only is
3178 useful in shell pipelines, and has little use in general
3179 code. This can be disabled with IgnoreSIPIPE=no in unit
3182 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
3183 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
3188 * This is mostly a bugfix release
3190 * We now expose the reason why a service failed in the
3191 "Result" D-Bus property.
3193 * Rudimentary service watchdog support (will be completed over
3194 the next few releases.)
3196 * When systemd forks off in order execute some service we will
3197 now immediately changes its argv[0] to reflect which process
3198 it will execute. This is useful to minimize the time window
3199 with a generic argv[0], which makes bootcharts more useful
3201 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
3202 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
3203 Mike Kazantsev, Ray Strode
3207 * This is mostly a test release, but incorporates many
3210 * New systemd-cgtop tool to show control groups by their
3213 * Linking against libacl for ACLs is optional again. If
3214 disabled, support tracking device access for active logins
3215 goes becomes unavailable, and so does access to the user
3216 journals by the respective users.
3218 * If a group "adm" exists, journal files are automatically
3219 owned by them, thus allow members of this group full access
3220 to the system journal as well as all user journals.
3222 * The journal now stores the SELinux context of the logging
3223 client for all entries.
3225 * Add C++ inclusion guards to all public headers
3227 * New output mode "cat" in the journal to print only text
3228 messages, without any meta data like date or time.
3230 * Include tiny X server wrapper as a temporary stop-gap to
3231 teach XOrg udev display enumeration. This is used by display
3232 managers such as gdm, and will go away as soon as XOrg
3233 learned native udev hotplugging for display devices.
3235 * Add new systemd-cat tool for executing arbitrary programs
3236 with STDERR/STDOUT connected to the journal. Can also act as
3237 BSD logger replacement, and does so by default.
3239 * Optionally store all locally generated coredumps in the
3240 journal along with meta data.
3242 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
3243 writing short strings to files (for usage for /sys), and for
3244 creating symlinks, character and block device nodes.
3246 * New unit file option ControlGroupPersistent= to make cgroups
3247 persistent, following the mechanisms outlined in
3248 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
3250 * Support multiple local RTCs in a sane way
3252 * No longer monopolize IO when replaying readahead data on
3253 rotating disks, since we might starve non-file-system IO to
3254 death, since fanotify() will not see accesses done by blkid,
3257 * Do not show kernel threads in systemd-cgls anymore, unless
3258 requested with new -k switch.
3260 Contributions from: Dan Horák, Kay Sievers, Lennart
3261 Poettering, Michal Schmidt
3265 * This is mostly a test release, but incorporates many
3268 * The git repository moved to:
3269 git://anongit.freedesktop.org/systemd/systemd
3270 ssh://git.freedesktop.org/git/systemd/systemd
3272 * First release with the journal
3273 http://0pointer.de/blog/projects/the-journal.html
3275 * The journal replaces both systemd-kmsg-syslogd and
3276 systemd-stdout-bridge.
3278 * New sd_pid_get_unit() API call in libsystemd-logind
3280 * Many systemadm clean-ups
3282 * Introduce remote-fs-pre.target which is ordered before all
3283 remote mounts and may be used to start services before all
3286 * Added Mageia support
3288 * Add bash completion for systemd-loginctl
3290 * Actively monitor PID file creation for daemons which exit in
3291 the parent process before having finished writing the PID
3292 file in the daemon process. Daemons which do this need to be
3293 fixed (i.e. PID file creation must have finished before the
3294 parent exits), but we now react a bit more gracefully to them.
3296 * Add colourful boot output, mimicking the well-known output
3297 of existing distributions.
3299 * New option PassCredentials= for socket units, for
3300 compatibility with a recent kernel ABI breakage.
3302 * /etc/rc.local is now hooked in via a generator binary, and
3303 thus will no longer act as synchronization point during
3306 * systemctl list-unit-files now supports --root=.
3308 * systemd-tmpfiles now understands two new commands: z, Z for
3309 relabelling files according to the SELinux database. This is
3310 useful to apply SELinux labels to specific files in /sys,
3313 * Output of SysV services is now forwarded to both the console
3314 and the journal by default, not only just the console.
3316 * New man pages for all APIs from libsystemd-login.
3318 * The build tree got reorganized and a the build system is a
3319 lot more modular allowing embedded setups to specifically
3320 select the components of systemd they are interested in.
3322 * Support for Linux systems lacking the kernel VT subsystem is
3325 * configure's --with-rootdir= got renamed to
3326 --with-rootprefix= to follow the naming used by udev and
3329 * Unless specified otherwise we will now install to /usr instead
3330 of /usr/local by default.
3332 * Processes with '@' in argv[0][0] are now excluded from the
3333 final shut-down killing spree, following the logic explained
3335 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
3337 * All processes remaining in a service cgroup when we enter
3338 the START or START_PRE states are now killed with
3339 SIGKILL. That means it is no longer possible to spawn
3340 background processes from ExecStart= lines (which was never
3341 supported anyway, and bad style).
3343 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
3344 reloading of units together.
3346 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
3347 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
3348 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
3349 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
3350 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
~ianmdlvl / elogind.git / blob