1 systemd System and Service Manager
5 * As an experimental feature, udev now tries to lock the
6 disk device node (flock(LOCK_SH|LOCK_NB)) while it
7 executes events for the disk or any of its partitions.
8 Applications like partitioning programs can lock the
9 disk device node (flock(LOCK_EX)) and claim temporary
10 device ownership that way; udev will entirely skip all event
11 handling for this disk and its partitions. If the disk
12 was opened for writing, the close will trigger a partition
13 table rescan in udev's "watch" facility, and if needed
14 synthesize "change" events for the disk and all its partitions.
15 This is now unconditionally enabled, if it turns out to
16 cause major problems, we might turn it on only for specific
17 devices, or might need to disable it entirely. Device-mapper
18 devices are excluded from this logic.
20 * We temporarily dropped the "-l" switch for fsck invocations,
21 since they collide with the flock() logic above. util-linux
22 upstream has been changed already to avoid this conflict,
23 and we will readd "-l" as soon as util-linux with this
24 change has been released.
26 * The dependency on libattr has been removed. Since a long
27 time the extended attribute calls have moved to glibc, and
28 libattr is thus unnecessary.
30 * Virtualization detection works without priviliges now. This
31 means the systemd-detect-virt binary no longer requires
32 CAP_SYS_PTRACE file capabilities, and our daemons can run
33 with fewer privileges.
35 * systemd-networkd now runs under its own "systemd-network"
36 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
37 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
38 loses the ability to write to files owned by root this way.
40 * Similar, systemd-resolved now runs under its own
41 "systemd-resolve" user with no capabilities remaining.
43 * Similar, systemd-bus-proxyd now runs under its own
44 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
46 * systemd-networkd gained support for setting up "veth"
47 virtual ethernet devices for container connectivity, as well
48 as GRE and VTI tunnels.
50 * systemd-networkd will no longer automatically attempt to
51 manually load kernel modules necessary for certain tunnel
52 transports. Instead it is assumed the kernel loads them
53 automatically when required. This only works correctly on
54 very new kernels. On older kernels, please consider adding
55 the kernel modules to /etc/load-modules.d/ as a work-around.
57 * The resolv.conf file systemd-resolved generates has been
58 moved to /run/systemd/resolve/, if you have a symlink from
59 /etc/resolv.conf it might be necessary to correct it.
61 * Two new service settings ProtectedHome= and ProtectedSystem=
62 have been added. When enabled they will make the user data
63 (such as /home) inaccessible or read-only and the system
64 (such as /usr) read-only, for specific services. This allows
65 very light-weight per-service sandboxing to avoid
66 modifications of user data or system files from
67 services. These two new switches have been enabled for all
68 of systemd's long-running services, where appropriate.
70 * Socket units gained new SocketUser= and SocketGroup=
71 settings to set the owner user and group of AF_UNIX sockets
72 and FIFOs in the file system.
74 * Socket units gained a new RemoveOnStop= setting. If enabled
75 all FIFOS and sockets in the file system will be removed
76 when the specific socket unit is stopped.
78 * Socket units gained a new Symlinks= setting. It takes a list
79 of symlinks to create to file system sockets or FIFOs
80 created by the specific unix sockets. This is useful to
81 manage symlinks to socket nodes with the same life-cycle as
84 * The /dev/log socket and /dev/initctl FIFO have been moved to
85 /run, and have been replaced by symlinks. This allows
86 connecting to these facilities even if PrivateDevices=yes is
87 used for a service (which makes /dev/log itself unavailable,
88 but /run is left). This also has the benefit of ensuring
89 that /dev only contains device nodes, directories and
90 symlinks, and nothing else.
92 * sd-daemon gained two new calls sd_pid_notify() and
93 sd_pid_notifyf(). They are similar to sd_notify() and
94 sd_notifyf(), but allow overriding of the source PID of
95 notification messages if permissions permit this. This is
96 useful to send notify messages on behalf of a different
97 process (for example, the parent process). The
98 systemd-notify tool has been updated to make use of this
99 when sending messages (so that notification messages now
100 originate from the shell script invoking systemd-notify and
101 not the systemd-notify process itself. This should minimize
102 a race where systemd fails to associate notification
103 messages to services when the originating process already
106 * A new "on-abnormal" setting for Restart= has been added. If
107 set it will result in automatic restarts on all "abnormal"
108 reasons for a process to exit, which includes unclean
109 signals, core dumps, timeouts and watchdog timeouts, but
110 does not include clean and unclean exit codes or clean
111 signals. Restart=on-abnormal is an alternative for
112 Restart=on-failure for services that shall be able to
113 terminate and avoid restarts on certain errors, by
114 indicating so with an unclean exit code. Restart=on-failure
115 or Restart=on-abnormal is now the recommended setting for
116 all long-running services.
118 * If the InaccessibleDirectories= service setting points to a
119 mount point (or if there are any submounts contained within
120 it), it is now attempted to completely unmount it, to make
121 the file systems truly unavailable for the respective
124 * The ReadOnlyDirectories= service setting and
125 systemd-nspawn's --read-only parameter are now recursively
126 applied to all submounts, too.
128 * Mount units may now be created transiently via the bus APIs.
130 * The support for SysV and LSB init scripts has been removed
131 from the systemd daemon itself. Instead, it is now
132 implemented as a generator that creates native systemd units
133 from these scripts when needed. This enables us to remove a
134 substantial amount of legacy code from PID 1, following the
135 fact that many distributions only ship a very small number
136 of LSB/SysV init scripts nowadays.
138 * Priviliged Xen (dom0) domains are not considered
139 virtualization anymore by the virtualization detection
140 logic. After all, they generally have unrestricted access to
141 the hardware and usually are used to manage the unprivileged
144 * systemd-tmpfiles gained a new "C" line type, for copying
145 files or entire directories.
147 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
148 lines. So far they have been non-globbing versions of the
149 latter, and have thus been redundant. In future it is
150 recommended to only use "z"; and "m" has hence been removed
151 from the documentation, even though it stays supported.
153 * A tmpfiles snippet to recreate the most basic structure in
154 /var has been added. This is enough to create the /var/run →
155 /run symlink and create a couple of structural
156 directories. This allows systems to boot up with an empty or
157 volatile /var. Of course, while with this change the core OS
158 now is capable with dealing with a volatile /var not all
159 user services are ready for it. However, we hope that sooner
160 or later many service daemons will be changed upstream so
161 that they are able to automatically create their necessary
162 directories in /var at boot, should they be missing. This is
163 the first step to allow state-less systems that only require
164 the vendor image for /usr to boot.
166 * systemd-nspawn has gained a new --tmpfs= switch to mount an
167 empty tmpfs instance to a specific directory. This is
168 particularly useful for making use of the automatic
169 reconstruction of /var (see above), by passing --tmpfs=/var.
171 * Access modes specified in tmpfiles snippets may now be
172 prefixed with "~", which indicates that they shall be masked
173 by whether the existing file or directly is currently
174 writable, readable or executable at all. Also, if specified
175 the sgid/suid/sticky bits will be masked for all
178 * A new passive target unit "network-pre.target" has been
179 added which is useful for services that shall run before any
180 network is configured, for example firewall scripts.
182 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
183 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
184 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
185 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
186 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
187 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
190 -- Berlin, 2014-06-11
194 * A new "systemd-timesyncd" daemon has been added for
195 synchronizing the system clock across the network. It
196 implements an SNTP client. In contrast to NTP
197 implementations such as chrony or the NTP reference server
198 this only implements a client side, and does not bother with
199 the full NTP complexity, focusing only on querying time from
200 one remote server and synchronizing the local clock to
201 it. Unless you intend to serve NTP to networked clients or
202 want to connect to local hardware clocks this simple NTP
203 client should be more than appropriate for most
204 installations. The daemon runs with minimal privileges, and
205 has been hooked up with networkd to only operate when
206 network connectivity is available. The daemon saves the
207 current clock to disk every time a new NTP sync has been
208 acquired, and uses this to possibly correct the system clock
209 early at bootup, in order to accommodate for systems that
210 lack an RTC such as the Raspberry Pi and embedded devices,
211 and make sure that time monotonically progresses on these
212 systems, even if it is not always correct. To make use of
213 this daemon a new system user and group "systemd-timesync"
214 needs to be created on installation of systemd.
216 * The queue "seqnum" interface of libudev has been disabled, as
217 it was generally incompatible with device namespacing as
218 sequence numbers of devices go "missing" if the devices are
219 part of a different namespace.
221 * "systemctl list-timers" and "systemctl list-sockets" gained
222 a --recursive switch for showing units of these types also
223 for all local containers, similar in style to the already
224 supported --recursive switch for "systemctl list-units".
226 * A new RebootArgument= setting has been added for service
227 units, which may be used to specify a kernel reboot argument
228 to use when triggering reboots with StartLimitAction=.
230 * A new FailureAction= setting has been added for service
231 units which may be used to specify an operation to trigger
232 when a service fails. This works similarly to
233 StartLimitAction=, but unlike it controls what is done
234 immediately rather than only after several attempts to
235 restart the service in question.
237 * hostnamed got updated to also expose the kernel name,
238 release, and version on the bus. This is useful for
239 executing commands like hostnamectl with the -H switch.
240 systemd-analyze makes use of this to properly display
241 details when running non-locally.
243 * The bootchart tool can now show cgroup information in the
246 * The CFS CPU quota cgroup attribute is now exposed for
247 services. The new CPUQuota= switch has been added for this
248 which takes a percentage value. Setting this will have the
249 result that a service may never get more CPU time than the
250 specified percentage, even if the machine is otherwise idle.
252 * systemd-networkd learned IPIP and SIT tunnel support.
254 * LSB init scripts exposing a dependency on $network will now
255 get a dependency on network-online.target rather than simply
256 network.target. This should bring LSB handling closer to
257 what it was on SysV systems.
259 * A new fsck.repair= kernel option has been added to control
260 how fsck shall deal with unclean file systems at boot.
262 * The (.ini) configuration file parser will now silently
263 ignore sections whose name begins with "X-". This may be
264 used to maintain application-specific extension sections in unit
267 * machined gained a new API to query the IP addresses of
268 registered containers. "machinectl status" has been updated
269 to show these addresses in its output.
271 * A new call sd_uid_get_display() has been added to the
272 sd-login APIs for querying the "primary" session of a
273 user. The "primary" session of the user is elected from the
274 user's sessions and generally a graphical session is
275 preferred over a text one.
277 * A minimal systemd-resolved daemon has been added. It
278 currently simply acts as a companion to systemd-networkd and
279 manages resolv.conf based on per-interface DNS
280 configuration, possibly supplied via DHCP. In the long run
281 we hope to extend this into a local DNSSEC enabled DNS and
284 * The systemd-networkd-wait-online tool is now enabled by
285 default. It will delay network-online.target until a network
286 connection has been configured. The tool primarily integrates
287 with networkd, but will also make a best effort to make sense
288 of network configuration performed in some other way.
290 * Two new service options StartupCPUShares= and
291 StartupBlockIOWeight= have been added that work similarly to
292 CPUShares= and BlockIOWeight= however only apply during
293 system startup. This is useful to prioritize certain services
294 differently during bootup than during normal runtime.
296 * hostnamed has been changed to prefer the statically
297 configured hostname in /etc/hostname (unless set to
298 'localhost' or empty) over any dynamic one supplied by
299 dhcp. With this change the rules for picking the hostname
300 match more closely the rules of other configuration settings
301 where the local administrator's configuration in /etc always
302 overrides any other settings.
304 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
305 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
306 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
307 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
308 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
309 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
310 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
311 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
312 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
313 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
314 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
315 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
316 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
317 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
318 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
319 Lindskog, WaLyong Cho, Will Woods, Zbigniew
322 -- Beijing, 2014-05-28
326 * When restoring the screen brightness at boot, stay away from
327 the darkest setting or from the lowest 5% of the available
328 range, depending on which is the larger value of both. This
329 should effectively protect the user from rebooting into a
330 black screen, should the brightness have been set to minimum
333 * sd-login gained a new sd_machine_get_class() call to
334 determine the class ("vm" or "container") of a machine
335 registered with machined.
337 * sd-login gained new calls
338 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
339 to query the identity of the peer of a local AF_UNIX
340 connection. They operate similarly to their sd_pid_get_xyz()
343 * PID 1 will now maintain a system-wide system state engine
344 with the states "starting", "running", "degraded",
345 "maintenance", "stopping". These states are bound to system
346 startup, normal runtime, runtime with at least one failed
347 service, rescue/emergency mode and system shutdown. This
348 state is shown in the "systemctl status" output when no unit
349 name is passed. It is useful to determine system state, in
350 particularly when doing so for many systems or containers at
353 * A new command "list-machines" has been added to "systemctl"
354 that lists all local OS containers and shows their system
355 state (see above), if systemd runs inside of them.
357 * systemctl gained a new "-r" switch to recursively enumerate
358 units on all local containers, when used with the
359 "list-unit" command (which is the default one that is
360 executed when no parameters are specified).
362 * The GPT automatic partition discovery logic will now honour
363 two GPT partition flags: one may be set on a partition to
364 cause it to be mounted read-only, and the other may be set
365 on a partition to ignore it during automatic discovery.
367 * Two new GPT type UUIDs have been added for automatic root
368 partition discovery, for 32-bit and 64-bit ARM. This is not
369 particularly useful for discovering the root directory on
370 these architectures during bare-metal boots (since UEFI is
371 not common there), but still very useful to allow booting of
372 ARM disk images in nspawn with the -i option.
374 * MAC addresses of interfaces created with nspawn's
375 --network-interface= switch will now be generated from the
376 machine name, and thus be stable between multiple invocations
379 * logind will now automatically remove all IPC objects owned
380 by a user if she or he fully logs out. This makes sure that
381 users who are logged out cannot continue to consume IPC
382 resources. This covers SysV memory, semaphores and message
383 queues as well as POSIX shared memory and message
384 queues. Traditionally, SysV and POSIX IPC had no life-cycle
385 limits. With this functionality, that is corrected. This may
386 be turned off by using the RemoveIPC= switch of logind.conf.
388 * The systemd-machine-id-setup and tmpfiles tools gained a
389 --root= switch to operate on a specific root directory,
392 * journald can now forward logged messages to the TTYs of all
393 logged in users ("wall"). This is the default for all
394 emergency messages now.
396 * A new tool systemd-journal-remote has been added to stream
397 journal log messages across the network.
399 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
400 controller trees are mounted into it. Note that the
401 directories mounted beneath it are not read-only. This is a
402 security measure and is particularly useful because glibc
403 actually includes a search logic to pick any tmpfs it can
404 find to implement shm_open() if /dev/shm is not available
405 (which it might very well be in namespaced setups).
407 * machinectl gained a new "poweroff" command to cleanly power
408 down a local OS container.
410 * The PrivateDevices= unit file setting will now also drop the
411 CAP_MKNOD capability from the capability bound set, and
412 imply DevicePolicy=closed.
414 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
415 comprehensively on all long-running systemd services where
418 * systemd-udevd will now run in a disassociated mount
419 namespace. To mount directories from udev rules, make sure to
420 pull in mount units via SYSTEMD_WANTS properties.
422 * The kdbus support gained support for uploading policy into
423 the kernel. sd-bus gained support for creating "monitoring"
424 connections that can eavesdrop into all bus communication
425 for debugging purposes.
427 * Timestamps may now be specified in seconds since the UNIX
428 epoch Jan 1st, 1970 by specifying "@" followed by the value
431 * Native tcpwrap support in systemd has been removed. tcpwrap
432 is old code, not really maintained anymore and has serious
433 shortcomings, and better options such as firewalls
434 exist. For setups that require tcpwrap usage, please
435 consider invoking your socket-activated service via tcpd,
436 like on traditional inetd.
438 * A new system.conf configuration option
439 DefaultTimerAccuracySec= has been added that controls the
440 default AccuracySec= setting of .timer units.
442 * Timer units gained a new WakeSystem= switch. If enabled,
443 timers configured this way will cause the system to resume
444 from system suspend (if the system supports that, which most
447 * Timer units gained a new Persistent= switch. If enabled,
448 timers configured this way will save to disk when they have
449 been last triggered. This information is then used on next
450 reboot to possible execute overdue timer events, that
451 could not take place because the system was powered off.
452 This enables simple anacron-like behaviour for timer units.
454 * systemctl's "list-timers" will now also list the time a
455 timer unit was last triggered in addition to the next time
456 it will be triggered.
458 * systemd-networkd will now assign predictable IPv4LL
459 addresses to its local interfaces.
461 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
462 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
463 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
464 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
465 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
466 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
467 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
468 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
471 -- Berlin, 2014-03-25
475 * A new unit file setting RestrictAddressFamilies= has been
476 added to restrict which socket address families unit
477 processes gain access to. This takes address family names
478 like "AF_INET" or "AF_UNIX", and is useful to minimize the
479 attack surface of services via exotic protocol stacks. This
480 is built on seccomp system call filters.
482 * Two new unit file settings RuntimeDirectory= and
483 RuntimeDirectoryMode= have been added that may be used to
484 manage a per-daemon runtime directories below /run. This is
485 an alternative for setting up directory permissions with
486 tmpfiles snippets, and has the advantage that the runtime
487 directory's lifetime is bound to the daemon runtime and that
488 the daemon starts up with an empty directory each time. This
489 is particularly useful when writing services that drop
490 privileges using the User= or Group= setting.
492 * The DeviceAllow= unit setting now supports globbing for
493 matching against device group names.
495 * The systemd configuration file system.conf gained new
496 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
497 DefaultMemoryAccounting= to globally turn on/off accounting
498 for specific resources (cgroups) for all units. These
499 settings may still be overridden individually in each unit
502 * systemd-gpt-auto-generator is now able to discover /srv and
503 root partitions in addition to /home and swap partitions. It
504 also supports LUKS-encrypted partitions now. With this in
505 place, automatic discovery of partitions to mount following
506 the Discoverable Partitions Specification
507 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
508 is now a lot more complete. This allows booting without
509 /etc/fstab and without root= on the kernel command line on
510 systems prepared appropriately.
512 * systemd-nspawn gained a new --image= switch which allows
513 booting up disk images and Linux installations on any block
514 device that follow the Discoverable Partitions Specification
515 (see above). This means that installations made with
516 appropriately updated installers may now be started and
517 deployed using container managers, completely
518 unmodified. (We hope that libvirt-lxc will add support for
519 this feature soon, too.)
521 * systemd-nspawn gained a new --network-macvlan= setting to
522 set up a private macvlan interface for the
523 container. Similarly, systemd-networkd gained a new
524 Kind=macvlan setting in .netdev files.
526 * systemd-networkd now supports configuring local addresses
529 * A new tool systemd-network-wait-online has been added to
530 synchronously wait for network connectivity using
533 * The sd-bus.h bus API gained a new sd_bus_track object for
534 tracking the life-cycle of bus peers. Note that sd-bus.h is
535 still not a public API though (unless you specify
536 --enable-kdbus on the configure command line, which however
537 voids your warranty and you get no API stability guarantee).
539 * The $XDG_RUNTIME_DIR runtime directories for each user are
540 now individual tmpfs instances, which has the benefit of
541 introducing separate pools for each user, with individual
542 size limits, and thus making sure that unprivileged clients
543 can no longer negatively impact the system or other users by
544 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
545 RuntimeDirectorySize= has been introduced that allows
546 controlling the default size limit for all users. It
547 defaults to 10% of the available physical memory. This is no
548 replacement for quotas on tmpfs though (which the kernel
549 still does not support), as /dev/shm and /tmp are still
550 shared resources used by both the system and unprivileged
553 * logind will now automatically turn off automatic suspending
554 on laptop lid close when more than one display is
555 connected. This was previously expected to be implemented
556 individually in desktop environments (such as GNOME),
557 however has been added to logind now, in order to fix a
558 boot-time race where a desktop environment might not have
559 been started yet and thus not been able to take an inhibitor
560 lock at the time where logind already suspends the system
563 * logind will now wait at least 30s after each system
564 suspend/resume cycle, and 3min after system boot before
565 suspending the system due to a closed laptop lid. This
566 should give USB docking stations and similar enough time to
567 be probed and configured after system resume and boot in
568 order to then act as suspend blocker.
570 * systemd-run gained a new --property= setting which allows
571 initialization of resource control properties (and others)
572 for the created scope or service unit. Example: "systemd-run
573 --property=BlockIOWeight=10 updatedb" may be used to run
574 updatedb at a low block IO scheduling weight.
576 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
577 now also work in --scope mode.
579 * When systemd is compiled with kdbus support, basic support
580 for enforced policies is now in place. (Note that enabling
581 kdbus still voids your warranty and no API compatibility
584 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
585 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
586 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
587 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
588 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
589 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
590 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
591 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
592 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
593 Zbigniew Jędrzejewski-Szmek
595 -- Berlin, 2014-03-12
599 * systemd will now relabel /dev after loading the SMACK policy
600 according to SMACK rules.
602 * A new unit file option AppArmorProfile= has been added to
603 set the AppArmor profile for the processes of a unit.
605 * A new condition check ConditionArchitecture= has been added
606 to conditionalize units based on the system architecture, as
607 reported by uname()'s "machine" field.
609 * systemd-networkd now supports matching on the system
610 virtualization, architecture, kernel command line, host name
613 * logind is now a lot more aggressive when suspending the
614 machine due to a closed laptop lid. Instead of acting only
615 on the lid close action, it will continuously watch the lid
616 status and act on it. This is useful for laptops where the
617 power button is on the outside of the chassis so that it can
618 be reached without opening the lid (such as the Lenovo
619 Yoga). On those machines, logind will now immediately
620 re-suspend the machine if the power button has been
621 accidentally pressed while the laptop was suspended and in a
624 * logind will now watch SW_DOCK switches and inhibit reaction
625 to the lid switch if it is pressed. This means that logind
626 will not suspend the machine anymore if the lid is closed
627 and the system is docked, if the laptop supports SW_DOCK
628 notifications via the input layer. Note that ACPI docking
629 stations do not generate this currently. Also note that this
630 logic is usually not fully sufficient and Desktop
631 Environments should take a lid switch inhibitor lock when an
632 external display is connected, as systemd will not watch
635 * nspawn will now make use of the devices cgroup controller by
636 default, and only permit creation of and access to the usual
637 API device nodes like /dev/null or /dev/random, as well as
638 access to (but not creation of) the pty devices.
640 * We will now ship a default .network file for
641 systemd-networkd that automatically configures DHCP for
642 network interfaces created by nspawn's --network-veth or
643 --network-bridge= switches.
645 * systemd will now understand the usual M, K, G, T suffixes
646 according to SI conventions (i.e. to the base 1000) when
647 referring to throughput and hardware metrics. It will stay
648 with IEC conventions (i.e. to the base 1024) for software
649 metrics, according to what is customary according to
650 Wikipedia. We explicitly document which base applies for
651 each configuration option.
653 * The DeviceAllow= setting in unit files now supports a syntax
654 to whitelist an entire group of devices node majors at once,
655 based on the /proc/devices listing. For example, with the
656 string "char-pts", it is now possible to whitelist all
657 current and future pseudo-TTYs at once.
659 * sd-event learned a new "post" event source. Event sources of
660 this type are triggered by the dispatching of any event
661 source of a type that is not "post". This is useful for
662 implementing clean-up and check event sources that are
663 triggered by other work being done in the program.
665 * systemd-networkd is no longer statically enabled, but uses
666 the usual [Install] sections so that it can be
667 enabled/disabled using systemctl. It still is enabled by
670 * When creating a veth interface pair with systemd-nspawn, the
671 host side will now be prefixed with "vb-" if
672 --network-bridge= is used, and with "ve-" if --network-veth
673 is used. This way, it is easy to distinguish these cases on
674 the host, for example to apply different configuration to
675 them with systemd-networkd.
677 * The compatibility libraries for libsystemd-journal.so,
678 libsystem-id128.so, libsystemd-login.so and
679 libsystemd-daemon.so do not make use of IFUNC
680 anymore. Instead, we now build libsystemd.so multiple times
681 under these alternative names. This means that the footprint
682 is drastically increased, but given that these are
683 transitional compatibility libraries, this should not matter
684 much. This change has been made necessary to support the ARM
685 platform for these compatibility libraries, as the ARM
686 toolchain is not really at the same level as the toolchain
687 for other architectures like x86 and does not support
688 IFUNC. Please make sure to use --enable-compat-libs only
689 during a transitional period!
691 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
692 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
693 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
694 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
695 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
696 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
697 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
698 Zbigniew Jędrzejewski-Szmek
700 -- Berlin, 2014-02-24
704 * A new component "systemd-networkd" has been added that can
705 be used to configure local network interfaces statically or
706 via DHCP. It is capable of bringing up bridges, VLANs, and
707 bonding. Currently, no hook-ups for interactive network
708 configuration are provided. Use this for your initrd,
709 container, embedded, or server setup if you need a simple,
710 yet powerful, network configuration solution. This
711 configuration subsystem is quite nifty, as it allows wildcard
712 hotplug matching in interfaces. For example, with a single
713 configuration snippet, you can configure that all Ethernet
714 interfaces showing up are automatically added to a bridge,
715 or similar. It supports link-sensing and more.
717 * A new tool "systemd-socket-proxyd" has been added which can
718 act as a bidirectional proxy for TCP sockets. This is
719 useful for adding socket activation support to services that
720 do not actually support socket activation, including virtual
721 machines and the like.
723 * Add a new tool to save/restore rfkill state on
726 * Save/restore state of keyboard backlights in addition to
727 display backlights on shutdown/boot.
729 * udev learned a new SECLABEL{} construct to label device
730 nodes with a specific security label when they appear. For
731 now, only SECLABEL{selinux} is supported, but the syntax is
732 prepared for additional security frameworks.
734 * udev gained a new scheme to configure link-level attributes
735 from files in /etc/systemd/network/*.link. These files can
736 match against MAC address, device path, driver name and type,
737 and will apply attributes like the naming policy, link speed,
738 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
739 address assignment policy (randomized, ...).
741 * The configuration of network interface naming rules for
742 "permanent interface names" has changed: a new NamePolicy=
743 setting in the [Link] section of .link files determines the
744 priority of possible naming schemes (onboard, slot, mac,
745 path). The default value of this setting is determined by
746 /usr/lib/net/links/99-default.link. Old
747 80-net-name-slot.rules udev configuration file has been
748 removed, so local configuration overriding this file should
749 be adapated to override 99-default.link instead.
751 * When the User= switch is used in a unit file, also
752 initialize $SHELL= based on the user database entry.
754 * systemd no longer depends on libdbus. All communication is
755 now done with sd-bus, systemd's low-level bus library
758 * kdbus support has been added to PID 1 itself. When kdbus is
759 enabled, this causes PID 1 to set up the system bus and
760 enable support for a new ".busname" unit type that
761 encapsulates bus name activation on kdbus. It works a little
762 bit like ".socket" units, except for bus names. A new
763 generator has been added that converts classic dbus1 service
764 activation files automatically into native systemd .busname
767 * sd-bus: add a light-weight vtable implementation that allows
768 defining objects on the bus with a simple static const
769 vtable array of its methods, signals and properties.
771 * systemd will not generate or install static dbus
772 introspection data anymore to /usr/share/dbus-1/interfaces,
773 as the precise format of these files is unclear, and
774 nothing makes use of it.
776 * A proxy daemon is now provided to proxy clients connecting
777 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
778 compatibility with classic D-Bus.
780 * A bus driver implementation has been added that supports the
781 classic D-Bus bus driver calls on kdbus, also for
782 compatibility purposes.
784 * A new API "sd-event.h" has been added that implements a
785 minimal event loop API built around epoll. It provides a
786 couple of features that direct epoll usage is lacking:
787 prioritization of events, scales to large numbers of timer
788 events, per-event timer slack (accuracy), system-wide
789 coalescing of timer events, exit handlers, watchdog
790 supervision support using systemd's sd_notify() API, child
793 * A new API "sd-rntl.h" has been added that provides an API
794 around the route netlink interface of the kernel, similar in
797 * A new API "sd-dhcp-client.h" has been added that provides a
798 small DHCPv4 client-side implementation. This is used by
801 * There is a new kernel command line option
802 "systemd.restore_state=0|1". When set to "0", none of the
803 systemd tools will restore saved runtime state to hardware
804 devices. More specifically, the rfkill and backlight states
807 * The FsckPassNo= compatibility option in mount/service units
808 has been removed. The fstab generator will now add the
809 necessary dependencies automatically, and does not require
810 PID1's support for that anymore.
812 * journalctl gained a new switch, --list-boots, that lists
813 recent boots with their times and boot IDs.
815 * The various tools like systemctl, loginctl, timedatectl,
816 busctl, systemd-run, ... have gained a new switch "-M" to
817 connect to a specific, local OS container (as direct
818 connection, without requiring SSH). This works on any
819 container that is registered with machined, such as those
820 created by libvirt-lxc or nspawn.
822 * systemd-run and systemd-analyze also gained support for "-H"
823 to connect to remote hosts via SSH. This is particularly
824 useful for systemd-run because it enables queuing of jobs
827 * machinectl gained a new command "login" to open a getty
828 login in any local container. This works with any container
829 that is registered with machined (such as those created by
830 libvirt-lxc or nspawn), and which runs systemd inside.
832 * machinectl gained a new "reboot" command that may be used to
833 trigger a reboot on a specific container that is registered
834 with machined. This works on any container that runs an init
837 * systemctl gained a new "list-timers" command to print a nice
838 listing of installed timer units with the times they elapse
841 * Alternative reboot() parameters may now be specified on the
842 "systemctl reboot" command line and are passed to the
843 reboot() system call.
845 * systemctl gained a new --job-mode= switch to configure the
846 mode to queue a job with. This is a more generic version of
847 --fail, --irreversible, and --ignore-dependencies, which are
848 still available but not advertised anymore.
850 * /etc/systemd/system.conf gained new settings to configure
851 various default timeouts of units, as well as the default
852 start limit interval and burst. These may still be overridden
855 * PID1 will now export on the bus profile data of the security
856 policy upload process (such as the SELinux policy upload to
859 * journald: when forwarding logs to the console, include
860 timestamps (following the setting in
861 /sys/module/printk/parameters/time).
863 * OnCalendar= in timer units now understands the special
864 strings "yearly" and "annually". (Both are equivalent)
866 * The accuracy of timer units is now configurable with the new
867 AccuracySec= setting. It defaults to 1min.
869 * A new dependency type JoinsNamespaceOf= has been added that
870 allows running two services within the same /tmp and network
871 namespace, if PrivateNetwork= or PrivateTmp= are used.
873 * A new command "cat" has been added to systemctl. It outputs
874 the original unit file of a unit, and concatenates the
875 contents of additional "drop-in" unit file snippets, so that
876 the full configuration is shown.
878 * systemctl now supports globbing on the various "list-xyz"
879 commands, like "list-units" or "list-sockets", as well as on
880 those commands which take multiple unit names.
882 * journalctl's --unit= switch gained support for globbing.
884 * All systemd daemons now make use of the watchdog logic so
885 that systemd automatically notices when they hang.
887 * If the $container_ttys environment variable is set,
888 getty-generator will automatically spawn a getty for each
889 listed tty. This is useful for container managers to request
890 login gettys to be spawned on as many ttys as needed.
892 * %h, %s, %U specifier support is not available anymore when
893 used in unit files for PID 1. This is because NSS calls are
894 not safe from PID 1. They stay available for --user
895 instances of systemd, and as special case for the root user.
897 * loginctl gained a new "--no-legend" switch to turn off output
900 * The "sd-login.h" API gained three new calls:
901 sd_session_is_remote(), sd_session_get_remote_user(),
902 sd_session_get_remote_host() to query information about
905 * The udev hardware database now also carries vendor/product
906 information of SDIO devices.
908 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
909 determine whether watchdog notifications are requested by
912 * Socket-activated per-connection services now include a
913 short description of the connection parameters in the
916 * tmpfiles gained a new "--boot" option. When this is not used,
917 only lines where the command character is not suffixed with
918 "!" are executed. When this option is specified, those
919 options are executed too. This partitions tmpfiles
920 directives into those that can be safely executed at any
921 time, and those which should be run only at boot (for
922 example, a line that creates /run/nologin).
924 * A new API "sd-resolve.h" has been added which provides a simple
925 asynchronous wrapper around glibc NSS host name resolution
926 calls, such as getaddrinfo(). In contrast to glibc's
927 getaddrinfo_a(), it does not use signals. In contrast to most
928 other asynchronous name resolution libraries, this one does
929 not reimplement DNS, but reuses NSS, so that alternate
930 host name resolution systems continue to work, such as mDNS,
931 LDAP, etc. This API is based on libasyncns, but it has been
932 cleaned up for inclusion in systemd.
934 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
935 "sd-daemon.h" are no longer found in individual libraries
936 libsystemd-journal.so, libsystemd-login.so,
937 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
938 merged them into a single library, libsystemd.so, which
939 provides all symbols. The reason for this is cyclic
940 dependencies, as these libraries tend to use each other's
941 symbols. So far, we have managed to workaround that by linking
942 a copy of a good part of our code into each of these
943 libraries again and again, which, however, makes certain
944 things hard to do, like sharing static variables. Also, it
945 substantially increases footprint. With this change, there
946 is only one library for the basic APIs systemd
947 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
948 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
949 library as well, however are subject to the --enable-kdbus
950 switch (see below). Note that "sd-dhcp-client.h" is not part
951 of this library (this is because it only consumes, never
952 provides, services of/to other APIs). To make the transition
953 easy from the separate libraries to the unified one, we
954 provide the --enable-compat-libs compile-time switch which
955 will generate stub libraries that are compatible with the
956 old ones but redirect all calls to the new one.
958 * All of the kdbus logic and the new APIs "sd-bus.h",
959 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
960 and "sd-utf8.h" are compile-time optional via the
961 "--enable-kdbus" switch, and they are not compiled in by
962 default. To make use of kdbus, you have to explicitly enable
963 the switch. Note however, that neither the kernel nor the
964 userspace API for all of this is considered stable yet. We
965 want to maintain the freedom to still change the APIs for
966 now. By specifying this build-time switch, you acknowledge
967 that you are aware of the instability of the current
970 * Also, note that while kdbus is pretty much complete,
971 it lacks one thing: proper policy support. This means you
972 can build a fully working system with all features; however,
973 it will be highly insecure. Policy support will be added in
974 one of the next releases, at the same time that we will
975 declare the APIs stable.
977 * When the kernel command-line argument "kdbus" is specified,
978 systemd will automatically load the kdbus.ko kernel module. At
979 this stage of development, it is only useful for testing kdbus
980 and should not be used in production. Note: if "--enable-kdbus"
981 is specified, and the kdbus.ko kernel module is available, and
982 "kdbus" is added to the kernel command line, the entire system
983 runs with kdbus instead of dbus-daemon, with the above mentioned
984 problem of missing the system policy enforcement. Also a future
985 version of kdbus.ko or a newer systemd will not be compatible with
986 each other, and will unlikely be able to boot the machine if only
987 one of them is updated.
989 * systemctl gained a new "import-environment" command which
990 uploads the caller's environment (or parts thereof) into the
991 service manager so that it is inherited by services started
992 by the manager. This is useful to upload variables like
993 $DISPLAY into the user service manager.
995 * A new PrivateDevices= switch has been added to service units
996 which allows running a service with a namespaced /dev
997 directory that does not contain any device nodes for
998 physical devices. More specifically, it only includes devices
999 such as /dev/null, /dev/urandom, and /dev/zero which are API
1002 * logind has been extended to support behaviour like VT
1003 switching on seats that do not support a VT. This makes
1004 multi-session available on seats that are not the first seat
1005 (seat0), and on systems where kernel support for VTs has
1006 been disabled at compile-time.
1008 * If a process holds a delay lock for system sleep or shutdown
1009 and fails to release it in time, we will now log its
1010 identity. This makes it easier to identify processes that
1011 cause slow suspends or power-offs.
1013 * When parsing /etc/crypttab, support for a new key-slot=
1014 option as supported by Debian is added. It allows indicating
1015 which LUKS slot to use on disk, speeding up key loading.
1017 * The sd_journald_sendv() API call has been checked and
1018 officially declared to be async-signal-safe so that it may
1019 be invoked from signal handlers for logging purposes.
1021 * Boot-time status output is now enabled automatically after a
1022 short timeout if boot does not progress, in order to give
1023 the user an indication what she or he is waiting for.
1025 * The boot-time output has been improved to show how much time
1026 remains until jobs expire.
1028 * The KillMode= switch in service units gained a new possible
1029 value "mixed". If set, and the unit is shut down, then the
1030 initial SIGTERM signal is sent only to the main daemon
1031 process, while the following SIGKILL signal is sent to
1032 all remaining processes of the service.
1034 * When a scope unit is registered, a new property "Controller"
1035 may be set. If set to a valid bus name, systemd will send a
1036 RequestStop() signal to this name when it would like to shut
1037 down the scope. This may be used to hook manager logic into
1038 the shutdown logic of scope units. Also, scope units may now
1039 be put in a special "abandoned" state, in which case the
1040 manager process which created them takes no further
1041 responsibilities for it.
1043 * When reading unit files, systemd will now verify
1044 the access mode of these files, and warn about certain
1045 suspicious combinations. This has been added to make it
1046 easier to track down packaging bugs where unit files are
1047 marked executable or world-writable.
1049 * systemd-nspawn gained a new "--setenv=" switch to set
1050 container-wide environment variables. The similar option in
1051 systemd-activate was renamed from "--environment=" to
1052 "--setenv=" for consistency.
1054 * systemd-nspawn has been updated to create a new kdbus domain
1055 for each container that is invoked, thus allowing each
1056 container to have its own set of system and user buses,
1057 independent of the host.
1059 * systemd-nspawn gained a new --drop-capability= switch to run
1060 the container with less capabilities than the default. Both
1061 --drop-capability= and --capability= now take the special
1062 string "all" for dropping or keeping all capabilities.
1064 * systemd-nspawn gained new switches for executing containers
1065 with specific SELinux labels set.
1067 * systemd-nspawn gained a new --quiet switch to not generate
1068 any additional output but the container's own console
1071 * systemd-nspawn gained a new --share-system switch to run a
1072 container without PID namespacing enabled.
1074 * systemd-nspawn gained a new --register= switch to control
1075 whether the container is registered with systemd-machined or
1076 not. This is useful for containers that do not run full
1077 OS images, but only specific apps.
1079 * systemd-nspawn gained a new --keep-unit which may be used
1080 when invoked as the only program from a service unit, and
1081 results in registration of the unit service itself in
1082 systemd-machined, instead of a newly opened scope unit.
1084 * systemd-nspawn gained a new --network-interface= switch for
1085 moving arbitrary interfaces to the container. The new
1086 --network-veth switch creates a virtual Ethernet connection
1087 between host and container. The new --network-bridge=
1088 switch then allows assigning the host side of this virtual
1089 Ethernet connection to a bridge device.
1091 * systemd-nspawn gained a new --personality= switch for
1092 setting the kernel personality for the container. This is
1093 useful when running a 32-bit container on a 64-bit host. A
1094 similar option Personality= is now also available for service
1097 * logind will now also track a "Desktop" identifier for each
1098 session which encodes the desktop environment of it. This is
1099 useful for desktop environments that want to identify
1100 multiple running sessions of itself easily.
1102 * A new SELinuxContext= setting for service units has been
1103 added that allows setting a specific SELinux execution
1104 context for a service.
1106 * Most systemd client tools will now honour $SYSTEMD_LESS for
1107 settings of the "less" pager. By default, these tools will
1108 override $LESS to allow certain operations to work, such as
1109 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
1110 influence this logic.
1112 * systemd's "seccomp" hook-up has been changed to make use of
1113 the libseccomp library instead of using its own
1114 implementation. This has benefits for portability among
1117 * For usage together with SystemCallFilter=, a new
1118 SystemCallErrorNumber= setting has been introduced that
1119 allows configuration of a system error number to be returned
1120 on filtered system calls, instead of immediately killing the
1121 process. Also, SystemCallArchitectures= has been added to
1122 limit access to system calls of a particular architecture
1123 (in order to turn off support for unused secondary
1124 architectures). There is also a global
1125 SystemCallArchitectures= setting in system.conf now to turn
1126 off support for non-native system calls system-wide.
1128 * systemd requires a kernel with a working name_to_handle_at(),
1129 please see the kernel config requirements in the README file.
1131 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
1132 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
1133 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
1134 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
1135 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
1136 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
1137 Elia Pinto, Florian Weimer, George McCollister, Goffredo
1138 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
1139 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
1140 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
1141 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
1142 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
1143 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
1144 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
1145 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
1146 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
1147 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
1148 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
1149 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
1150 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
1151 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
1152 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
1153 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
1154 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
1156 -- Berlin, 2014-02-20
1160 * logind has gained support for facilitating privileged input
1161 and drm device access for unprivileged clients. This work is
1162 useful to allow Wayland display servers (and similar
1163 programs, such as kmscon) to run under the user's ID and
1164 access input and drm devices which are normally
1165 protected. When this is used (and the kernel is new enough)
1166 logind will "mute" IO on the file descriptors passed to
1167 Wayland as long as it is in the background and "unmute" it
1168 if it returns into the foreground. This allows secure
1169 session switching without allowing background sessions to
1170 eavesdrop on input and display data. This also introduces
1171 session switching support if VT support is turned off in the
1172 kernel, and on seats that are not seat0.
1174 * A new kernel command line option luks.options= is understood
1175 now which allows specifiying LUKS options for usage for LUKS
1176 encrypted partitions specified with luks.uuid=.
1178 * tmpfiles.d(5) snippets may now use specifier expansion in
1179 path names. More specifically %m, %b, %H, %v, are now
1180 replaced by the local machine id, boot id, hostname, and
1181 kernel version number.
1183 * A new tmpfiles.d(5) command "m" has been introduced which
1184 may be used to change the owner/group/access mode of a file
1185 or directory if it exists, but do nothing if it does not.
1187 * This release removes high-level support for the
1188 MemorySoftLimit= cgroup setting. The underlying kernel
1189 cgroup attribute memory.soft_limit= is currently badly
1190 designed and likely to be removed from the kernel API in its
1191 current form, hence we should not expose it for now.
1193 * The memory.use_hierarchy cgroup attribute is now enabled for
1194 all cgroups systemd creates in the memory cgroup
1195 hierarchy. This option is likely to be come the built-in
1196 default in the kernel anyway, and the non-hierarchial mode
1197 never made much sense in the intrinsically hierarchial
1200 * A new field _SYSTEMD_SLICE= is logged along with all journal
1201 messages containing the slice a message was generated
1202 from. This is useful to allow easy per-customer filtering of
1203 logs among other things.
1205 * systemd-journald will no longer adjust the group of journal
1206 files it creates to the "systemd-journal" group. Instead we
1207 rely on the journal directory to be owned by the
1208 "systemd-journal" group, and its setgid bit set, so that the
1209 kernel file system layer will automatically enforce that
1210 journal files inherit this group assignment. The reason for
1211 this change is that we cannot allow NSS look-ups from
1212 journald which would be necessary to resolve
1213 "systemd-journal" to a numeric GID, because this might
1214 create deadlocks if NSS involves synchronous queries to
1215 other daemons (such as nscd, or sssd) which in turn are
1216 logging clients of journald and might block on it, which
1217 would then dead lock. A tmpfiles.d(5) snippet included in
1218 systemd will make sure the setgid bit and group are
1219 properly set on the journal directory if it exists on every
1220 boot. However, we recommend adjusting it manually after
1221 upgrades too (or from RPM scriptlets), so that the change is
1222 not delayed until next reboot.
1224 * Backlight and random seed files in /var/lib/ have moved into
1225 the /var/lib/systemd/ directory, in order to centralize all
1226 systemd generated files in one directory.
1228 * Boot time performance measurements (as displayed by
1229 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
1230 performance information if that's available to determine how
1231 much time BIOS and boot loader initialization required. With
1232 a sufficiently new BIOS you hence no longer need to boot
1233 with Gummiboot to get access to such information.
1235 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
1236 Cristian Rodríguez, Dave Reisner, David Herrmann, David
1237 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
1238 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
1239 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
1240 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
1241 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1243 -- Berlin, 2013-10-02
1247 * The Restart= option for services now understands a new
1248 on-watchdog setting, which will restart the service
1249 automatically if the service stops sending out watchdog keep
1250 alive messages (as configured with WatchdogSec=).
1252 * The getty generator (which is responsible for bringing up a
1253 getty on configured serial consoles) will no longer only
1254 start a getty on the primary kernel console but on all
1255 others, too. This makes the order in which console= is
1256 specified on the kernel command line less important.
1258 * libsystemd-logind gained a new sd_session_get_vt() call to
1259 retrieve the VT number of a session.
1261 * If the option "tries=0" is set for an entry of /etc/crypttab
1262 its passphrase is queried indefinitely instead of any
1263 maximum number of tries.
1265 * If a service with a configure PID file terminates its PID
1266 file will now be removed automatically if it still exists
1267 afterwards. This should put an end to stale PID files.
1269 * systemd-run will now also take relative binary path names
1270 for execution and no longer insists on absolute paths.
1272 * InaccessibleDirectories= and ReadOnlyDirectories= now take
1273 paths that are optionally prefixed with "-" to indicate that
1274 it should not be considered a failure if they do not exist.
1276 * journalctl -o (and similar commands) now understands a new
1277 output mode "short-precise", it is similar to "short" but
1278 shows timestamps with usec accuracy.
1280 * The option "discard" (as known from Debian) is now
1281 synonymous to "allow-discards" in /etc/crypttab. In fact,
1282 "discard" is preferred now (since it is easier to remember
1285 * Some licensing clean-ups were made, so that more code is now
1286 LGPL-2.1 licensed than before.
1288 * A minimal tool to save/restore the display backlight
1289 brightness across reboots has been added. It will store the
1290 backlight setting as late as possible at shutdown, and
1291 restore it as early as possible during reboot.
1293 * A logic to automatically discover and enable home and swap
1294 partitions on GPT disks has been added. With this in place
1295 /etc/fstab becomes optional for many setups as systemd can
1296 discover certain partitions located on the root disk
1297 automatically. Home partitions are recognized under their
1298 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
1299 partitions are recognized under their GPT type ID
1300 0657fd6da4ab43c484e50933c84b4f4f.
1302 * systemd will no longer pass any environment from the kernel
1303 or initrd to system services. If you want to set an
1304 environment for all services, do so via the kernel command
1305 line systemd.setenv= assignment.
1307 * The systemd-sysctl tool no longer natively reads the file
1308 /etc/sysctl.conf. If desired, the file should be symlinked
1309 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
1310 legacy support by a symlink rather than built-in code, it
1311 also makes the otherwise hidden order of application of the
1312 different files visible. (Note that this partly reverts to a
1313 pre-198 application order of sysctl knobs!)
1315 * The "systemctl set-log-level" and "systemctl dump" commands
1316 have been moved to systemd-analyze.
1318 * systemd-run learned the new --remain-after-exit switch,
1319 which causes the scope unit not to be cleaned up
1320 automatically after the process terminated.
1322 * tmpfiles learned a new --exclude-prefix= switch to exclude
1323 certain paths from operation.
1325 * journald will now automatically flush all messages to disk
1326 as soon as a message of the log priorities CRIT, ALERT or
1329 Contributions from: Andrew Cook, Brandon Philips, Christian
1330 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
1331 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
1332 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
1333 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
1334 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
1335 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
1336 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
1337 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
1338 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
1339 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
1340 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
1341 William Giokas, Zbigniew Jędrzejewski-Szmek
1343 -- Berlin, 2013-09-13
1347 * The documentation has been updated to cover the various new
1348 concepts introduced with 205.
1350 * Unit files now understand the new %v specifier which
1351 resolves to the kernel version string as returned by "uname
1354 * systemctl now supports filtering the unit list output by
1355 load state, active state and sub state, using the new
1358 * "systemctl status" will now show the results of the
1359 condition checks (like ConditionPathExists= and similar) of
1360 the last start attempts of the unit. They are also logged to
1363 * "journalctl -b" may now be used to look for boot output of a
1364 specific boot. Try "journalctl -b -1" for the previous boot,
1365 but the syntax is substantially more powerful.
1367 * "journalctl --show-cursor" has been added which prints the
1368 cursor string the last shown log line. This may then be used
1369 with the new "journalctl --after-cursor=" switch to continue
1370 browsing logs from that point on.
1372 * "journalctl --force" may now be used to force regeneration
1375 * Creation of "dead" device nodes has been moved from udev
1376 into kmod and tmpfiles. Previously, udev would read the kmod
1377 databases to pre-generate dead device nodes based on meta
1378 information contained in kernel modules, so that these would
1379 be auto-loaded on access rather then at boot. As this
1380 does not really have much to do with the exposing actual
1381 kernel devices to userspace this has always been slightly
1382 alien in the udev codebase. Following the new scheme kmod
1383 will now generate a runtime snippet for tmpfiles from the
1384 module meta information and it now is tmpfiles' job to the
1385 create the nodes. This also allows overriding access and
1386 other parameters for the nodes using the usual tmpfiles
1387 facilities. As side effect this allows us to remove the
1388 CAP_SYS_MKNOD capability bit from udevd entirely.
1390 * logind's device ACLs may now be applied to these "dead"
1391 devices nodes too, thus finally allowing managed access to
1392 devices such as /dev/snd/sequencer whithout loading the
1393 backing module right-away.
1395 * A new RPM macro has been added that may be used to apply
1396 tmpfiles configuration during package installation.
1398 * systemd-detect-virt and ConditionVirtualization= now can
1399 detect User-Mode-Linux machines (UML).
1401 * journald will now implicitly log the effective capabilities
1402 set of processes in the message metadata.
1404 * systemd-cryptsetup has gained support for TrueCrypt volumes.
1406 * The initrd interface has been simplified (more specifically,
1407 support for passing performance data via environment
1408 variables and fsck results via files in /run has been
1409 removed). These features were non-essential, and are
1410 nowadays available in a much nicer way by having systemd in
1411 the initrd serialize its state and have the hosts systemd
1412 deserialize it again.
1414 * The udev "keymap" data files and tools to apply keyboard
1415 specific mappings of scan to key codes, and force-release
1416 scan code lists have been entirely replaced by a udev
1417 "keyboard" builtin and a hwdb data file.
1419 * systemd will now honour the kernel's "quiet" command line
1420 argument also during late shutdown, resulting in a
1421 completely silent shutdown when used.
1423 * There's now an option to control the SO_REUSEPORT socket
1424 option in .socket units.
1426 * Instance units will now automatically get a per-template
1427 subslice of system.slice unless something else is explicitly
1428 configured. For example, instances of sshd@.service will now
1429 implicitly be placed in system-sshd.slice rather than
1430 system.slice as before.
1432 * Test coverage support may now be enabled at build time.
1434 Contributions from: Dave Reisner, Frederic Crozat, Harald
1435 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
1436 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
1437 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
1438 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
1439 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
1440 Giokas, Zbigniew Jędrzejewski-Szmek
1442 -- Berlin, 2013-07-23
1446 * Two new unit types have been introduced:
1448 Scope units are very similar to service units, however, are
1449 created out of pre-existing processes -- instead of PID 1
1450 forking off the processes. By using scope units it is
1451 possible for system services and applications to group their
1452 own child processes (worker processes) in a powerful way
1453 which then maybe used to organize them, or kill them
1454 together, or apply resource limits on them.
1456 Slice units may be used to partition system resources in an
1457 hierarchial fashion and then assign other units to them. By
1458 default there are now three slices: system.slice (for all
1459 system services), user.slice (for all user sessions),
1460 machine.slice (for VMs and containers).
1462 Slices and scopes have been introduced primarily in
1463 context of the work to move cgroup handling to a
1464 single-writer scheme, where only PID 1
1465 creates/removes/manages cgroups.
1467 * There's a new concept of "transient" units. In contrast to
1468 normal units these units are created via an API at runtime,
1469 not from configuration from disk. More specifically this
1470 means it is now possible to run arbitrary programs as
1471 independent services, with all execution parameters passed
1472 in via bus APIs rather than read from disk. Transient units
1473 make systemd substantially more dynamic then it ever was,
1474 and useful as a general batch manager.
1476 * logind has been updated to make use of scope and slice units
1477 for managing user sessions. As a user logs in he will get
1478 his own private slice unit, to which all sessions are added
1479 as scope units. We also added support for automatically
1480 adding an instance of user@.service for the user into the
1481 slice. Effectively logind will no longer create cgroup
1482 hierarchies on its own now, it will defer entirely to PID 1
1483 for this by means of scope, service and slice units. Since
1484 user sessions this way become entities managed by PID 1
1485 the output of "systemctl" is now a lot more comprehensive.
1487 * A new mini-daemon "systemd-machined" has been added which
1488 may be used by virtualization managers to register local
1489 VMs/containers. nspawn has been updated accordingly, and
1490 libvirt will be updated shortly. machined will collect a bit
1491 of meta information about the VMs/containers, and assign
1492 them their own scope unit (see above). The collected
1493 meta-data is then made available via the "machinectl" tool,
1494 and exposed in "ps" and similar tools. machined/machinectl
1495 is compile-time optional.
1497 * As discussed earlier, the low-level cgroup configuration
1498 options ControlGroup=, ControlGroupModify=,
1499 ControlGroupPersistent=, ControlGroupAttribute= have been
1500 removed. Please use high-level attribute settings instead as
1501 well as slice units.
1503 * A new bus call SetUnitProperties() has been added to alter
1504 various runtime parameters of a unit. This is primarily
1505 useful to alter cgroup parameters dynamically in a nice way,
1506 but will be extended later on to make more properties
1507 modifiable at runtime. systemctl gained a new set-properties
1508 command that wraps this call.
1510 * A new tool "systemd-run" has been added which can be used to
1511 run arbitrary command lines as transient services or scopes,
1512 while configuring a number of settings via the command
1513 line. This tool is currently very basic, however already
1514 very useful. We plan to extend this tool to even allow
1515 queuing of execution jobs with time triggers from the
1516 command line, similar in fashion to "at".
1518 * nspawn will now inform the user explicitly that kernels with
1519 audit enabled break containers, and suggest the user to turn
1522 * Support for detecting the IMA and AppArmor security
1523 frameworks with ConditionSecurity= has been added.
1525 * journalctl gained a new "-k" switch for showing only kernel
1526 messages, mimicking dmesg output; in addition to "--user"
1527 and "--system" switches for showing only user's own logs
1530 * systemd-delta can now show information about drop-in
1531 snippets extending unit files.
1533 * libsystemd-bus has been substantially updated but is still
1534 not available as public API.
1536 * systemd will now look for the "debug" argument on the kernel
1537 command line and enable debug logging, similar to what
1538 "systemd.log_level=debug" already did before.
1540 * "systemctl set-default", "systemctl get-default" has been
1541 added to configure the default.target symlink, which
1542 controls what to boot into by default.
1544 * "systemctl set-log-level" has been added as a convenient
1545 way to raise and lower systemd logging threshold.
1547 * "systemd-analyze plot" will now show the time the various
1548 generators needed for execution, as well as information
1549 about the unit file loading.
1551 * libsystemd-journal gained a new sd_journal_open_files() call
1552 for opening specific journal files. journactl also gained a
1553 new switch to expose this new functionality. Previously we
1554 only supported opening all files from a directory, or all
1555 files from the system, as opening individual files only is
1556 racy due to journal file rotation.
1558 * systemd gained the new DefaultEnvironment= setting in
1559 /etc/systemd/system.conf to set environment variables for
1562 * If a privileged process logs a journal message with the
1563 OBJECT_PID= field set, then journald will automatically
1564 augment this with additional OBJECT_UID=, OBJECT_GID=,
1565 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
1566 system services want to log events about specific client
1567 processes. journactl/systemctl has been updated to make use
1568 of this information if all log messages regarding a specific
1571 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
1572 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
1573 Reisner, David Coppa, David King, David Strauss, Eelco
1574 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
1575 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
1576 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
1577 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
1578 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
1579 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
1580 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
1581 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
1582 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
1583 Łukasz Stelmach, 장동준
1587 * The Python bindings gained some minimal support for the APIs
1588 exposed by libsystemd-logind.
1590 * ConditionSecurity= gained support for detecting SMACK. Since
1591 this condition already supports SELinux and AppArmor we only
1592 miss IMA for this. Patches welcome!
1594 Contributions from: Karol Lewandowski, Lennart Poettering,
1595 Zbigniew Jędrzejewski-Szmek
1599 * systemd-nspawn will now create /etc/resolv.conf if
1600 necessary, before bind-mounting the host's file onto it.
1602 * systemd-nspawn will now store meta information about a
1603 container on the container's cgroup as extended attribute
1604 fields, including the root directory.
1606 * The cgroup hierarchy has been reworked in many ways. All
1607 objects any of the components systemd creates in the cgroup
1608 tree are now suffixed. More specifically, user sessions are
1609 now placed in cgroups suffixed with ".session", users in
1610 cgroups suffixed with ".user", and nspawn containers in
1611 cgroups suffixed with ".nspawn". Furthermore, all cgroup
1612 names are now escaped in a simple scheme to avoid collision
1613 of userspace object names with kernel filenames. This work
1614 is preparation for making these objects relocatable in the
1615 cgroup tree, in order to allow easy resource partitioning of
1616 these objects without causing naming conflicts.
1618 * systemctl list-dependencies gained the new switches
1619 --plain, --reverse, --after and --before.
1621 * systemd-inhibit now shows the process name of processes that
1622 have taken an inhibitor lock.
1624 * nss-myhostname will now also resolve "localhost"
1625 implicitly. This makes /etc/hosts an optional file and
1626 nicely handles that on IPv6 ::1 maps to both "localhost" and
1629 * libsystemd-logind.so gained a new call
1630 sd_get_machine_names() to enumerate running containers and
1631 VMs (currently only supported by very new libvirt and
1632 nspawn). sd_login_monitor can now be used to watch
1633 VMs/containers coming and going.
1635 * .include is not allowed recursively anymore, and only in
1636 unit files. Usually it is better to use drop-in snippets in
1637 .d/*.conf anyway, as introduced with systemd 198.
1639 * systemd-analyze gained a new "critical-chain" command that
1640 determines the slowest chain of units run during system
1641 boot-up. It is very useful for tracking down where
1642 optimizing boot time is the most beneficial.
1644 * systemd will no longer allow manipulating service paths in
1645 the name=systemd:/system cgroup tree using ControlGroup= in
1646 units. (But is still fine with it in all other dirs.)
1648 * There's a new systemd-nspawn@.service service file that may
1649 be used to easily run nspawn containers as system
1650 services. With the container's root directory in
1651 /var/lib/container/foobar it is now sufficient to run
1652 "systemctl start systemd-nspawn@foobar.service" to boot it.
1654 * systemd-cgls gained a new parameter "--machine" to list only
1655 the processes within a certain container.
1657 * ConditionSecurity= now can check for "apparmor". We still
1658 are lacking checks for SMACK and IMA for this condition
1659 check though. Patches welcome!
1661 * A new configuration file /etc/systemd/sleep.conf has been
1662 added that may be used to configure which kernel operation
1663 systemd is supposed to execute when "suspend", "hibernate"
1664 or "hybrid-sleep" is requested. This makes the new kernel
1665 "freeze" state accessible to the user.
1667 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
1668 the passed argument if applicable.
1670 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
1671 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
1672 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
1673 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
1674 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
1675 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
1676 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
1681 * The output of 'systemctl list-jobs' got some polishing. The
1682 '--type=' argument may now be passed more than once. A new
1683 command 'systemctl list-sockets' has been added which shows
1684 a list of kernel sockets systemd is listening on with the
1685 socket units they belong to, plus the units these socket
1688 * The experimental libsystemd-bus library got substantial
1689 updates to work in conjunction with the (also experimental)
1690 kdbus kernel project. It works well enough to exchange
1691 messages with some sophistication. Note that kdbus is not
1692 ready yet, and the library is mostly an elaborate test case
1693 for now, and not installable.
1695 * systemd gained a new unit 'systemd-static-nodes.service'
1696 that generates static device nodes earlier during boot, and
1697 can run in conjunction with udev.
1699 * libsystemd-login gained a new call sd_pid_get_user_unit()
1700 to retrieve the user systemd unit a process is running
1701 in. This is useful for systems where systemd is used as
1704 * systemd-nspawn now places all containers in the new /machine
1705 top-level cgroup directory in the name=systemd
1706 hierarchy. libvirt will soon do the same, so that we get a
1707 uniform separation of /system, /user and /machine for system
1708 services, user processes and containers/virtual
1709 machines. This new cgroup hierarchy is also useful to stick
1710 stable names to specific container instances, which can be
1711 recognized later this way (this name may be controlled
1712 via systemd-nspawn's new -M switch). libsystemd-login also
1713 gained a new call sd_pid_get_machine_name() to retrieve the
1714 name of the container/VM a specific process belongs to.
1716 * bootchart can now store its data in the journal.
1718 * libsystemd-journal gained a new call
1719 sd_journal_add_conjunction() for AND expressions to the
1720 matching logic. This can be used to express more complex
1721 logical expressions.
1723 * journactl can now take multiple --unit= and --user-unit=
1726 * The cryptsetup logic now understands the "luks.key=" kernel
1727 command line switch for specifying a file to read the
1728 decryption key from. Also, if a configured key file is not
1729 found the tool will now automatically fall back to prompting
1732 * Python systemd.journal module was updated to wrap recently
1733 added functions from libsystemd-journal. The interface was
1734 changed to bring the low level interface in s.j._Reader
1735 closer to the C API, and the high level interface in
1736 s.j.Reader was updated to wrap and convert all data about
1739 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
1740 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
1741 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
1742 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
1743 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
1744 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1748 * journalctl --update-catalog now understands a new --root=
1749 option to operate on catalogs found in a different root
1752 * During shutdown after systemd has terminated all running
1753 services a final killing loop kills all remaining left-over
1754 processes. We will now print the name of these processes
1755 when we send SIGKILL to them, since this usually indicates a
1758 * If /etc/crypttab refers to password files stored on
1759 configured mount points automatic dependencies will now be
1760 generated to ensure the specific mount is established first
1761 before the key file is attempted to be read.
1763 * 'systemctl status' will now show information about the
1764 network sockets a socket unit is listening on.
1766 * 'systemctl status' will also shown information about any
1767 drop-in configuration file for units. (Drop-In configuration
1768 files in this context are files such as
1769 /etc/systemd/systemd/foobar.service.d/*.conf)
1771 * systemd-cgtop now optionally shows summed up CPU times of
1772 cgroups. Press '%' while running cgtop to switch between
1773 percentage and absolute mode. This is useful to determine
1774 which cgroups use up the most CPU time over the entire
1775 runtime of the system. systemd-cgtop has also been updated
1776 to be 'pipeable' for processing with further shell tools.
1778 * 'hostnamectl set-hostname' will now allow setting of FQDN
1781 * The formatting and parsing of time span values has been
1782 changed. The parser now understands fractional expressions
1783 such as "5.5h". The formatter will now output fractional
1784 expressions for all time spans under 1min, i.e. "5.123456s"
1785 rather than "5s 123ms 456us". For time spans under 1s
1786 millisecond values are shown, for those under 1ms
1787 microsecond values are shown. This should greatly improve
1788 all time-related output of systemd.
1790 * libsystemd-login and libsystemd-journal gained new
1791 functions for querying the poll() events mask and poll()
1792 timeout value for integration into arbitrary event
1795 * localectl gained the ability to list available X11 keymaps
1796 (models, layouts, variants, options).
1798 * 'systemd-analyze dot' gained the ability to filter for
1799 specific units via shell-style globs, to create smaller,
1800 more useful graphs. I.e. it is now possible to create simple
1801 graphs of all the dependencies between only target units, or
1802 of all units that Avahi has dependencies with.
1804 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
1805 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
1806 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
1807 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
1808 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
1809 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
1810 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
1814 * The boot-time readahead implementation for rotating media
1815 will now read the read-ahead data in multiple passes which
1816 consist of all read requests made in equidistant time
1817 intervals. This means instead of strictly reading read-ahead
1818 data in its physical order on disk we now try to find a
1819 middle ground between physical and access time order.
1821 * /etc/os-release files gained a new BUILD_ID= field for usage
1822 on operating systems that provide continuous builds of OS
1825 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
1826 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
1827 William Douglas, Zbigniew Jędrzejewski-Szmek
1831 * systemd-python gained an API exposing libsystemd-daemon.
1833 * The SMACK setup logic gained support for uploading CIPSO
1836 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
1837 ReadOnlyDirectories= and InaccessibleDirectories= has
1838 changed. The private /tmp and /var/tmp directories are now
1839 shared by all processes of a service (which means
1840 ExecStartPre= may now leave data in /tmp that ExecStart= of
1841 the same service can still access). When a service is
1842 stopped its temporary directories are immediately deleted
1843 (normal clean-up with tmpfiles is still done in addition to
1846 * By default, systemd will now set a couple of sysctl
1847 variables in the kernel: the safe sysrq options are turned
1848 on, IP route verification is turned on, and source routing
1849 disabled. The recently added hardlink and softlink
1850 protection of the kernel is turned on. These settings should
1851 be reasonably safe, and good defaults for all new systems.
1853 * The predictable network naming logic may now be turned off
1854 with a new kernel command line switch: net.ifnames=0.
1856 * A new libsystemd-bus module has been added that implements a
1857 pretty complete D-Bus client library. For details see:
1859 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
1861 * journald will now explicitly flush the journal files to disk
1862 at the latest 5min after each write. The file will then also
1863 be marked offline until the next write. This should increase
1864 reliability in case of a crash. The synchronization delay
1865 can be configured via SyncIntervalSec= in journald.conf.
1867 * There's a new remote-fs-setup.target unit that can be used
1868 to pull in specific services when at least one remote file
1869 system is to be mounted.
1871 * There are new targets timers.target and paths.target as
1872 canonical targets to pull user timer and path units in
1873 from. This complements sockets.target with a similar
1874 purpose for socket units.
1876 * libudev gained a new call udev_device_set_attribute_value()
1877 to set sysfs attributes of a device.
1879 * The udev daemon now sets the default number of worker
1880 processes executed in parallel based on the number of available
1881 CPUs instead of the amount of available RAM. This is supposed
1882 to provide a more reliable default and limit a too aggressive
1883 paralellism for setups with 1000s of devices connected.
1885 Contributions from: Auke Kok, Colin Walters, Cristian
1886 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
1887 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
1888 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
1889 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
1890 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
1891 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
1892 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
1893 Zbigniew Jędrzejewski-Szmek
1897 * Configuration of unit files may now be extended via drop-in
1898 files without having to edit/override the unit files
1899 themselves. More specifically, if the administrator wants to
1900 change one value for a service file foobar.service he can
1901 now do so by dropping in a configuration snippet into
1902 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
1903 will load all these snippets and apply them on top of the
1904 main unit configuration file, possibly extending or
1905 overriding its settings. Using these drop-in snippets is
1906 generally nicer than the two earlier options for changing
1907 unit files locally: copying the files from
1908 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
1909 them there; or creating a new file in /etc/systemd/system/
1910 that incorporates the original one via ".include". Drop-in
1911 snippets into these .d/ directories can be placed in any
1912 directory systemd looks for units in, and the usual
1913 overriding semantics between /usr/lib, /etc and /run apply
1916 * Most unit file settings which take lists of items can now be
1917 reset by assigning the empty string to them. For example,
1918 normally, settings such as Environment=FOO=BAR append a new
1919 environment variable assignment to the environment block,
1920 each time they are used. By assigning Environment= the empty
1921 string the environment block can be reset to empty. This is
1922 particularly useful with the .d/*.conf drop-in snippets
1923 mentioned above, since this adds the ability to reset list
1924 settings from vendor unit files via these drop-ins.
1926 * systemctl gained a new "list-dependencies" command for
1927 listing the dependencies of a unit recursively.
1929 * Inhibitors are now honored and listed by "systemctl
1930 suspend", "systemctl poweroff" (and similar) too, not only
1931 GNOME. These commands will also list active sessions by
1934 * Resource limits (as exposed by the various control group
1935 controllers) can now be controlled dynamically at runtime
1936 for all units. More specifically, you can now use a command
1937 like "systemctl set-cgroup-attr foobar.service cpu.shares
1938 2000" to alter the CPU shares a specific service gets. These
1939 settings are stored persistently on disk, and thus allow the
1940 administrator to easily adjust the resource usage of
1941 services with a few simple commands. This dynamic resource
1942 management logic is also available to other programs via the
1943 bus. Almost any kernel cgroup attribute and controller is
1946 * systemd-vconsole-setup will now copy all font settings to
1947 all allocated VTs, where it previously applied them only to
1950 * libsystemd-login gained the new sd_session_get_tty() API
1953 * This release drops support for a few legacy or
1954 distribution-specific LSB facility names when parsing init
1955 scripts: $x-display-manager, $mail-transfer-agent,
1956 $mail-transport-agent, $mail-transfer-agent, $smtp,
1957 $null. Also, the mail-transfer-agent.target unit backing
1958 this has been removed. Distributions which want to retain
1959 compatibility with this should carry the burden for
1960 supporting this themselves and patch support for these back
1961 in, if they really need to. Also, the facilities $syslog and
1962 $local_fs are now ignored, since systemd does not support
1963 early-boot LSB init scripts anymore, and these facilities
1964 are implied anyway for normal services. syslog.target has
1967 * There are new bus calls on PID1's Manager object for
1968 cancelling jobs, and removing snapshot units. Previously,
1969 both calls were only available on the Job and Snapshot
1972 * systemd-journal-gatewayd gained SSL support.
1974 * The various "environment" files, such as /etc/locale.conf
1975 now support continuation lines with a backslash ("\") as
1976 last character in the line, similarly in style (but different)
1977 to how this is supported in shells.
1979 * For normal user processes the _SYSTEMD_USER_UNIT= field is
1980 now implicitly appended to every log entry logged. systemctl
1981 has been updated to filter by this field when operating on a
1982 user systemd instance.
1984 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
1985 CAP_AUDIT_CONTROL capabilities to the capabilities set for
1986 the container. This makes it easier to boot unmodified
1987 Fedora systems in a container, which however still requires
1988 audit=0 to be passed on the kernel command line. Auditing in
1989 kernel and userspace is unfortunately still too broken in
1990 context of containers, hence we recommend compiling it out
1991 of the kernel or using audit=0. Hopefully this will be fixed
1992 one day for good in the kernel.
1994 * nspawn gained the new --bind= and --bind-ro= parameters to
1995 bind mount specific directories from the host into the
1998 * nspawn will now mount its own devpts file system instance
1999 into the container, in order not to leak pty devices from
2000 the host into the container.
2002 * systemd will now read the firmware boot time performance
2003 information from the EFI variables, if the used boot loader
2004 supports this, and takes it into account for boot performance
2005 analysis via "systemd-analyze". This is currently supported
2006 only in conjunction with Gummiboot, but could be supported
2007 by other boot loaders too. For details see:
2009 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
2011 * A new generator has been added that automatically mounts the
2012 EFI System Partition (ESP) to /boot, if that directory
2013 exists, is empty, and no other file system has been
2014 configured to be mounted there.
2016 * logind will now send out PrepareForSleep(false) out
2017 unconditionally, after coming back from suspend. This may be
2018 used by applications as asynchronous notification for
2019 system resume events.
2021 * "systemctl unlock-sessions" has been added, that allows
2022 unlocking the screens of all user sessions at once, similar
2023 to how "systemctl lock-sessions" already locked all users
2024 sessions. This is backed by a new D-Bus call UnlockSessions().
2026 * "loginctl seat-status" will now show the master device of a
2027 seat. (i.e. the device of a seat that needs to be around for
2028 the seat to be considered available, usually the graphics
2031 * tmpfiles gained a new "X" line type, that allows
2032 configuration of files and directories (with wildcards) that
2033 shall be excluded from automatic cleanup ("aging").
2035 * udev default rules set the device node permissions now only
2036 at "add" events, and do not change them any longer with a
2037 later "change" event.
2039 * The log messages for lid events and power/sleep keypresses
2040 now carry a message ID.
2042 * We now have a substantially larger unit test suite, but this
2043 continues to be work in progress.
2045 * udevadm hwdb gained a new --root= parameter to change the
2046 root directory to operate relative to.
2048 * logind will now issue a background sync() request to the kernel
2049 early at shutdown, so that dirty buffers are flushed to disk early
2050 instead of at the last moment, in order to optimize shutdown
2053 * A new bootctl tool has been added that is an interface for
2054 certain boot loader operations. This is currently a preview
2055 and is likely to be extended into a small mechanism daemon
2056 like timedated, localed, hostnamed, and can be used by
2057 graphical UIs to enumerate available boot options, and
2058 request boot into firmware operations.
2060 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
2061 the rest of the package. It also has been updated to work
2062 correctly in initrds.
2064 * Policykit previously has been runtime optional, and is now
2065 also compile time optional via a configure switch.
2067 * systemd-analyze has been reimplemented in C. Also "systemctl
2068 dot" has moved into systemd-analyze.
2070 * "systemctl status" with no further parameters will now print
2071 the status of all active or failed units.
2073 * Operations such as "systemctl start" can now be executed
2074 with a new mode "--irreversible" which may be used to queue
2075 operations that cannot accidentally be reversed by a later
2076 job queuing. This is by default used to make shutdown
2077 requests more robust.
2079 * The Python API of systemd now gained a new module for
2080 reading journal files.
2082 * A new tool kernel-install has been added that can install
2083 kernel images according to the Boot Loader Specification:
2085 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
2087 * Boot time console output has been improved to provide
2088 animated boot time output for hanging jobs.
2090 * A new tool systemd-activate has been added which can be used
2091 to test socket activation with, directly from the command
2092 line. This should make it much easier to test and debug
2093 socket activation in daemons.
2095 * journalctl gained a new "--reverse" (or -r) option to show
2096 journal output in reverse order (i.e. newest line first).
2098 * journalctl gained a new "--pager-end" (or -e) option to jump
2099 to immediately jump to the end of the journal in the
2100 pager. This is only supported in conjunction with "less".
2102 * journalctl gained a new "--user-unit=" option, that works
2103 similarly to "--unit=" but filters for user units rather than
2106 * A number of unit files to ease adoption of systemd in
2107 initrds has been added. This moves some minimal logic from
2108 the various initrd implementations into systemd proper.
2110 * The journal files are now owned by a new group
2111 "systemd-journal", which exists specifically to allow access
2112 to the journal, and nothing else. Previously, we used the
2113 "adm" group for that, which however possibly covers more
2114 than just journal/log file access. This new group is now
2115 already used by systemd-journal-gatewayd to ensure this
2116 daemon gets access to the journal files and as little else
2117 as possible. Note that "make install" will also set FS ACLs
2118 up for /var/log/journal to give "adm" and "wheel" read
2119 access to it, in addition to "systemd-journal" which owns
2120 the journal files. We recommend that packaging scripts also
2121 add read access to "adm" + "wheel" to /var/log/journal, and
2122 all existing/future journal files. To normal users and
2123 administrators little changes, however packagers need to
2124 ensure to create the "systemd-journal" system group at
2125 package installation time.
2127 * The systemd-journal-gatewayd now runs as unprivileged user
2128 systemd-journal-gateway:systemd-journal-gateway. Packaging
2129 scripts need to create these system user/group at
2132 * timedated now exposes a new boolean property CanNTP that
2133 indicates whether a local NTP service is available or not.
2135 * systemd-detect-virt will now also detect xen PVs
2137 * The pstore file system is now mounted by default, if it is
2140 * In addition to the SELinux and IMA policies we will now also
2141 load SMACK policies at early boot.
2143 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
2144 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
2145 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
2146 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
2147 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
2148 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
2149 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
2150 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
2151 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
2152 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
2153 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
2154 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
2155 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
2156 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
2160 * Timer units now support calendar time events in addition to
2161 monotonic time events. That means you can now trigger a unit
2162 based on a calendar time specification such as "Thu,Fri
2163 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
2164 or fifth day of any month of the year 2013, given that it is
2165 a thursday or friday. This brings timer event support
2166 considerably closer to cron's capabilities. For details on
2167 the supported calendar time specification language see
2170 * udev now supports a number of different naming policies for
2171 network interfaces for predictable names, and a combination
2172 of these policies is now the default. Please see this wiki
2173 document for details:
2175 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
2177 * Auke Kok's bootchart implementation has been added to the
2178 systemd tree. It is an optional component that can graph the
2179 boot in quite some detail. It is one of the best bootchart
2180 implementations around and minimal in its code and
2183 * nss-myhostname has been integrated into the systemd source
2184 tree. nss-myhostname guarantees that the local hostname
2185 always stays resolvable via NSS. It has been a weak
2186 requirement of systemd-hostnamed since a long time, and
2187 since its code is actually trivial we decided to just
2188 include it in systemd's source tree. It can be turned off
2189 with a configure switch.
2191 * The read-ahead logic is now capable of properly detecting
2192 whether a btrfs file system is on SSD or rotating media, in
2193 order to optimize the read-ahead scheme. Previously, it was
2194 only capable of detecting this on traditional file systems
2197 * In udev, additional device properties are now read from the
2198 IAB in addition to the OUI database. Also, Bluetooth company
2199 identities are attached to the devices as well.
2201 * In service files %U may be used as specifier that is
2202 replaced by the configured user name of the service.
2204 * nspawn may now be invoked without a controlling TTY. This
2205 makes it suitable for invocation as its own service. This
2206 may be used to set up a simple containerized server system
2207 using only core OS tools.
2209 * systemd and nspawn can now accept socket file descriptors
2210 when they are started for socket activation. This enables
2211 implementation of socket activated nspawn
2212 containers. i.e. think about autospawning an entire OS image
2213 when the first SSH or HTTP connection is received. We expect
2214 that similar functionality will also be added to libvirt-lxc
2217 * journalctl will now suppress ANSI color codes when
2218 presenting log data.
2220 * systemctl will no longer show control group information for
2221 a unit if a the control group is empty anyway.
2223 * logind can now automatically suspend/hibernate/shutdown the
2226 * /etc/machine-info and hostnamed now also expose the chassis
2227 type of the system. This can be used to determine whether
2228 the local system is a laptop, desktop, handset or
2229 tablet. This information may either be configured by the
2230 user/vendor or is automatically determined from ACPI and DMI
2231 information if possible.
2233 * A number of PolicyKit actions are now bound together with
2234 "imply" rules. This should simplify creating UIs because
2235 many actions will now authenticate similar ones as well.
2237 * Unit files learnt a new condition ConditionACPower= which
2238 may be used to conditionalize a unit depending on whether an
2239 AC power source is connected or not, of whether the system
2240 is running on battery power.
2242 * systemctl gained a new "is-failed" verb that may be used in
2243 shell scripts and suchlike to check whether a specific unit
2244 is in the "failed" state.
2246 * The EnvironmentFile= setting in unit files now supports file
2247 globbing, and can hence be used to easily read a number of
2248 environment files at once.
2250 * systemd will no longer detect and recognize specific
2251 distributions. All distribution-specific #ifdeffery has been
2252 removed, systemd is now fully generic and
2253 distribution-agnostic. Effectively, not too much is lost as
2254 a lot of the code is still accessible via explicit configure
2255 switches. However, support for some distribution specific
2256 legacy configuration file formats has been dropped. We
2257 recommend distributions to simply adopt the configuration
2258 files everybody else uses now and convert the old
2259 configuration from packaging scripts. Most distributions
2260 already did that. If that's not possible or desirable,
2261 distributions are welcome to forward port the specific
2262 pieces of code locally from the git history.
2264 * When logging a message about a unit systemd will now always
2265 log the unit name in the message meta data.
2267 * localectl will now also discover system locale data that is
2268 not stored in locale archives, but directly unpacked.
2270 * logind will no longer unconditionally use framebuffer
2271 devices as seat masters, i.e. as devices that are required
2272 to be existing before a seat is considered preset. Instead,
2273 it will now look for all devices that are tagged as
2274 "seat-master" in udev. By default framebuffer devices will
2275 be marked as such, but depending on local systems other
2276 devices might be marked as well. This may be used to
2277 integrate graphics cards using closed source drivers (such
2278 as NVidia ones) more nicely into logind. Note however, that
2279 we recommend using the open source NVidia drivers instead,
2280 and no udev rules for the closed-source drivers will be
2281 shipped from us upstream.
2283 Contributions from: Adam Williamson, Alessandro Crismani, Auke
2284 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
2285 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
2286 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
2287 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
2288 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
2289 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
2290 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
2291 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
2292 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
2293 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
2294 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
2299 * udev gained support for loading additional device properties
2300 from an indexed database that is keyed by vendor/product IDs
2301 and similar device identifiers. For the beginning this
2302 "hwdb" is populated with data from the well-known PCI and
2303 USB database, but also includes PNP, ACPI and OID data. In
2304 the longer run this indexed database shall grow into
2305 becoming the one central database for non-essential
2306 userspace device metadata. Previously, data from the PCI/USB
2307 database was only attached to select devices, since the
2308 lookup was a relatively expensive operation due to O(n) time
2309 complexity (with n being the number of entries in the
2310 database). Since this is now O(1), we decided to add in this
2311 data for all devices where this is available, by
2312 default. Note that the indexed database needs to be rebuilt
2313 when new data files are installed. To achieve this you need
2314 to update your packaging scripts to invoke "udevadm hwdb
2315 --update" after installation of hwdb data files. For
2316 RPM-based distributions we introduced the new
2317 %udev_hwdb_update macro for this purpose.
2319 * The Journal gained support for the "Message Catalog", an
2320 indexed database to link up additional information with
2321 journal entries. For further details please check:
2323 http://www.freedesktop.org/wiki/Software/systemd/catalog
2325 The indexed message catalog database also needs to be
2326 rebuilt after installation of message catalog files. Use
2327 "journalctl --update-catalog" for this. For RPM-based
2328 distributions we introduced the %journal_catalog_update
2329 macro for this purpose.
2331 * The Python Journal bindings gained support for the standard
2332 Python logging framework.
2334 * The Journal API gained new functions for checking whether
2335 the underlying file system of a journal file is capable of
2336 properly reporting file change notifications, or whether
2337 applications that want to reflect journal changes "live"
2338 need to recheck journal files continuously in appropriate
2341 * It is now possible to set the "age" field for tmpfiles
2342 entries to 0, indicating that files matching this entry
2343 shall always be removed when the directories are cleaned up.
2345 * coredumpctl gained a new "gdb" verb which invokes gdb
2346 right-away on the selected coredump.
2348 * There's now support for "hybrid sleep" on kernels that
2349 support this, in addition to "suspend" and "hibernate". Use
2350 "systemctl hybrid-sleep" to make use of this.
2352 * logind's HandleSuspendKey= setting (and related settings)
2353 now gained support for a new "lock" setting to simply
2354 request the screen lock on all local sessions, instead of
2355 actually executing a suspend or hibernation.
2357 * systemd will now mount the EFI variables file system by
2360 * Socket units now gained support for configuration of the
2361 SMACK security label.
2363 * timedatectl will now output the time of the last and next
2364 daylight saving change.
2366 * We dropped support for various legacy and distro-specific
2367 concepts, such as insserv, early-boot SysV services
2368 (i.e. those for non-standard runlevels such as 'b' or 'S')
2369 or ArchLinux /etc/rc.conf support. We recommend the
2370 distributions who still need support this to either continue
2371 to maintain the necessary patches downstream, or find a
2372 different solution. (Talk to us if you have questions!)
2374 * Various systemd components will now bypass PolicyKit checks
2375 for root and otherwise handle properly if PolicyKit is not
2376 found to be around. This should fix most issues for
2377 PolicyKit-less systems. Quite frankly this should have been
2378 this way since day one. It is absolutely our intention to
2379 make systemd work fine on PolicyKit-less systems, and we
2380 consider it a bug if something does not work as it should if
2381 PolicyKit is not around.
2383 * For embedded systems it is now possible to build udev and
2384 systemd without blkid and/or kmod support.
2386 * "systemctl switch-root" is now capable of switching root
2387 more than once. I.e. in addition to transitions from the
2388 initrd to the host OS it is now possible to transition to
2389 further OS images from the host. This is useful to implement
2390 offline updating tools.
2392 * Various other additions have been made to the RPM macros
2393 shipped with systemd. Use %udev_rules_update() after
2394 installing new udev rules files. %_udevhwdbdir,
2395 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
2396 %_sysctldir are now available which resolve to the right
2397 directories for packages to place various data files in.
2399 * journalctl gained the new --full switch (in addition to
2400 --all, to disable ellipsation for long messages.
2402 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
2403 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
2404 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
2405 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
2406 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
2407 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
2408 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
2409 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
2410 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
2414 * journalctl gained new --since= and --until= switches to
2415 filter by time. It also now supports nice filtering for
2416 units via --unit=/-u.
2418 * Type=oneshot services may use ExecReload= and do the
2421 * The journal daemon now supports time-based rotation and
2422 vacuuming, in addition to the usual disk-space based
2425 * The journal will now index the available field values for
2426 each field name. This enables clients to show pretty drop
2427 downs of available match values when filtering. The bash
2428 completion of journalctl has been updated
2429 accordingly. journalctl gained a new switch -F to list all
2430 values a certain field takes in the journal database.
2432 * More service events are now written as structured messages
2433 to the journal, and made recognizable via message IDs.
2435 * The timedated, localed and hostnamed mini-services which
2436 previously only provided support for changing time, locale
2437 and hostname settings from graphical DEs such as GNOME now
2438 also have a minimal (but very useful) text-based client
2439 utility each. This is probably the nicest way to changing
2440 these settings from the command line now, especially since
2441 it lists available options and is fully integrated with bash
2444 * There's now a new tool "systemd-coredumpctl" to list and
2445 extract coredumps from the journal.
2447 * We now install a README each in /var/log/ and
2448 /etc/rc.d/init.d explaining where the system logs and init
2449 scripts went. This hopefully should help folks who go to
2450 that dirs and look into the otherwise now empty void and
2451 scratch their heads.
2453 * When user-services are invoked (by systemd --user) the
2454 $MANAGERPID env var is set to the PID of systemd.
2456 * SIGRTMIN+24 when sent to a --user instance will now result
2457 in immediate termination of systemd.
2459 * gatewayd received numerous feature additions such as a
2460 "follow" mode, for live syncing and filtering.
2462 * browse.html now allows filtering and showing detailed
2463 information on specific entries. Keyboard navigation and
2464 mouse screen support has been added.
2466 * gatewayd/journalctl now supports HTML5/JSON
2467 Server-Sent-Events as output.
2469 * The SysV init script compatibility logic will now
2470 heuristically determine whether a script supports the
2471 "reload" verb, and only then make this available as
2474 * "systemctl status --follow" has been removed, use "journalctl
2477 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
2478 have been removed since they are hardly useful to be
2481 * And I'd like to take the opportunity to specifically mention
2482 Zbigniew for his great contributions. Zbigniew, you rock!
2484 Contributions from: Andrew Eikum, Christian Hesse, Colin
2485 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
2486 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
2487 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
2488 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
2489 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
2490 Jędrzejewski-Szmek, Сковорода Никита Андреевич
2494 * If /etc/vconsole.conf is non-existent or empty we will no
2495 longer load any console font or key map at boot by
2496 default. Instead the kernel defaults will be left
2497 intact. This is definitely the right thing to do, as no
2498 configuration should mean no configuration, and hard-coding
2499 font names that are different on all archs is probably a bad
2500 idea. Also, the kernel default key map and font should be
2501 good enough for most cases anyway, and mostly identical to
2502 the userspace fonts/key maps we previously overloaded them
2503 with. If distributions want to continue to default to a
2504 non-kernel font or key map they should ship a default
2505 /etc/vconsole.conf with the appropriate contents.
2507 Contributions from: Colin Walters, Daniel J Walsh, Dave
2508 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
2509 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2513 * journalctl gained a new --cursor= switch to show entries
2514 starting from the specified location in the journal.
2516 * We now enforce a size limit on journal entry fields exported
2517 with "-o json" in journalctl. Fields larger than 4K will be
2518 assigned null. This can be turned off with --all.
2520 * An (optional) journal gateway daemon is now available as
2521 "systemd-journal-gatewayd.service". This service provides
2522 access to the journal via HTTP and JSON. This functionality
2523 will be used to implement live log synchronization in both
2524 pull and push modes, but has various other users too, such
2525 as easy log access for debugging of embedded devices. Right
2526 now it is already useful to retrieve the journal via HTTP:
2528 # systemctl start systemd-journal-gatewayd.service
2529 # wget http://localhost:19531/entries
2531 This will download the journal contents in a
2532 /var/log/messages compatible format. The same as JSON:
2534 # curl -H"Accept: application/json" http://localhost:19531/entries
2536 This service is also accessible via a web browser where a
2537 single static HTML5 app is served that uses the JSON logic
2538 to enable the user to do some basic browsing of the
2539 journal. This will be extended later on. Here's an example
2540 screenshot of this app in its current state:
2542 http://0pointer.de/public/journal-gatewayd
2544 Contributions from: Kay Sievers, Lennart Poettering, Robert
2545 Milasan, Tom Gundersen
2549 * The bash completion logic is now available for journalctl
2552 * We do not mount the "cpuset" controller anymore together with
2553 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
2554 started if no parameters are assigned to it. "cpuset" hence
2555 broke code that assumed it it could create "cpu" groups and
2558 * journalctl -f will now subscribe to terminal size changes,
2559 and line break accordingly.
2561 Contributions from: Dave Reisner, Kay Sievers, Lennart
2562 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
2566 * nspawn will now create a symlink /etc/localtime in the
2567 container environment, copying the host's timezone
2568 setting. Previously this has been done via a bind mount, but
2569 since symlinks cannot be bind mounted this has now been
2570 changed to create/update the appropriate symlink.
2572 * journalctl -n's line number argument is now optional, and
2573 will default to 10 if omitted.
2575 * journald will now log the maximum size the journal files may
2576 take up on disk. This is particularly useful if the default
2577 built-in logic of determining this parameter from the file
2578 system size is used. Use "systemctl status
2579 systemd-journald.service" to see this information.
2581 * The multi-seat X wrapper tool has been stripped down. As X
2582 is now capable of enumerating graphics devices via udev in a
2583 seat-aware way the wrapper is not strictly necessary
2584 anymore. A stripped down temporary stop-gap is still shipped
2585 until the upstream display managers have been updated to
2586 fully support the new X logic. Expect this wrapper to be
2587 removed entirely in one of the next releases.
2589 * HandleSleepKey= in logind.conf has been split up into
2590 HandleSuspendKey= and HandleHibernateKey=. The old setting
2591 is not available anymore. X11 and the kernel are
2592 distuingishing between these keys and we should too. This
2593 also means the inhibition lock for these keys has been split
2596 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
2597 Poettering, Lukas Nykryn, Václav Pavlín
2601 * Whenever a unit changes state we will now log this to the
2602 journal and show along the unit's own log output in
2605 * ConditionPathIsMountPoint= can now properly detect bind
2606 mount points too. (Previously, a bind mount of one file
2607 system to another place in the same file system could not be
2608 detected as mount, since they shared struct stat's st_dev
2611 * We will now mount the cgroup controllers cpu, cpuacct,
2612 cpuset and the controllers net_cls, net_prio together by
2615 * nspawn containers will now have a virtualized boot
2616 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
2617 over with a randomized ID at container initialization). This
2618 has the effect of making "journalctl -b" do the right thing
2621 * The JSON output journal serialization has been updated not
2622 to generate "endless" list objects anymore, but rather one
2623 JSON object per line. This is more in line how most JSON
2624 parsers expect JSON objects. The new output mode
2625 "json-pretty" has been added to provide similar output, but
2626 neatly aligned for readability by humans.
2628 * We dropped all explicit sync() invocations in the shutdown
2629 code. The kernel does this implicitly anyway in the kernel
2630 reboot() syscall. halt(8)'s -n option is now a compatibility
2633 * We now support virtualized reboot() in containers, as
2634 supported by newer kernels. We will fall back to exit() if
2635 CAP_SYS_REBOOT is not available to the container. Also,
2636 nspawn makes use of this now and will actually reboot the
2637 container if the containerized OS asks for that.
2639 * journalctl will only show local log output by default
2640 now. Use --merge (-m) to show remote log output, too.
2642 * libsystemd-journal gained the new sd_journal_get_usage()
2643 call to determine the current disk usage of all journal
2644 files. This is exposed in the new "journalctl --disk-usage"
2647 * journald gained a new configuration setting SplitMode= in
2648 journald.conf which may be used to control how user journals
2649 are split off. See journald.conf(5) for details.
2651 * A new condition type ConditionFileNotEmpty= has been added.
2653 * tmpfiles' "w" lines now support file globbing, to write
2654 multiple files at once.
2656 * We added Python bindings for the journal submission
2657 APIs. More Python APIs for a number of selected APIs will
2658 likely follow. Note that we intend to add native bindings
2659 only for the Python language, as we consider it common
2660 enough to deserve bindings shipped within systemd. There are
2661 various projects outside of systemd that provide bindings
2662 for languages such as PHP or Lua.
2664 * Many conditions will now resolve specifiers such as %i. In
2665 addition, PathChanged= and related directives of .path units
2666 now support specifiers as well.
2668 * There's now a new RPM macro definition for the system preset
2671 * journald will now warn if it ca not forward a message to the
2672 syslog daemon because its socket is full.
2674 * timedated will no longer write or process /etc/timezone,
2675 except on Debian. As we do not support late mounted /usr
2676 anymore /etc/localtime always being a symlink is now safe,
2677 and hence the information in /etc/timezone is not necessary
2680 * logind will now always reserve one VT for a text getty (VT6
2681 by default). Previously if more than 6 X sessions where
2682 started they took up all the VTs with auto-spawned gettys,
2683 so that no text gettys were available anymore.
2685 * udev will now automatically inform the btrfs kernel logic
2686 about btrfs RAID components showing up. This should make
2687 simple hotplug based btrfs RAID assembly work.
2689 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
2690 (but not for its children which will stay at the kernel
2691 default). This should allow setups with a lot more listening
2694 * systemd will now always pass the configured timezone to the
2695 kernel at boot. timedated will do the same when the timezone
2698 * logind's inhibition logic has been updated. By default,
2699 logind will now handle the lid switch, the power and sleep
2700 keys all the time, even in graphical sessions. If DEs want
2701 to handle these events on their own they should take the new
2702 handle-power-key, handle-sleep-key and handle-lid-switch
2703 inhibitors during their runtime. A simple way to achiveve
2704 that is to invoke the DE wrapped in an invocation of:
2706 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
2708 * Access to unit operations is now checked via SELinux taking
2709 the unit file label and client process label into account.
2711 * systemd will now notify the administrator in the journal
2712 when he over-mounts a non-empty directory.
2714 * There are new specifiers that are resolved in unit files,
2715 for the host name (%H), the machine ID (%m) and the boot ID
2718 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
2719 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
2720 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
2721 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
2722 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
2723 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2724 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
2728 * Support for reading structured kernel messages from
2729 /dev/kmsg has now been added and is enabled by default.
2731 * Support for reading kernel messages from /proc/kmsg has now
2732 been removed. If you want kernel messages in the journal
2733 make sure to run a recent kernel (>= 3.5) that supports
2734 reading structured messages from /dev/kmsg (see
2735 above). /proc/kmsg is now exclusive property of classic
2736 syslog daemons again.
2738 * The libudev API gained the new
2739 udev_device_new_from_device_id() call.
2741 * The logic for file system namespace (ReadOnlyDirectory=,
2742 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
2743 require pivot_root() anymore. This means fewer temporary
2744 directories are created below /tmp for this feature.
2746 * nspawn containers will now see and receive all submounts
2747 made on the host OS below the root file system of the
2750 * Forward Secure Sealing is now supported for Journal files,
2751 which provide cryptographical sealing of journal files so
2752 that attackers cannot alter log history anymore without this
2753 being detectable. Lennart will soon post a blog story about
2754 this explaining it in more detail.
2756 * There are two new service settings RestartPreventExitStatus=
2757 and SuccessExitStatus= which allow configuration of exit
2758 status (exit code or signal) which will be excepted from the
2759 restart logic, resp. consider successful.
2761 * journalctl gained the new --verify switch that can be used
2762 to check the integrity of the structure of journal files and
2763 (if Forward Secure Sealing is enabled) the contents of
2766 * nspawn containers will now be run with /dev/stdin, /dev/fd/
2767 and similar symlinks pre-created. This makes running shells
2768 as container init process a lot more fun.
2770 * The fstab support can now handle PARTUUID= and PARTLABEL=
2773 * A new ConditionHost= condition has been added to match
2774 against the hostname (with globs) and machine ID. This is
2775 useful for clusters where a single OS image is used to
2776 provision a large number of hosts which shall run slightly
2777 different sets of services.
2779 * Services which hit the restart limit will now be placed in a
2782 Contributions from: Bertram Poettering, Dave Reisner, Huang
2783 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
2784 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
2788 * When running in --user mode systemd will now become a
2789 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
2790 tree a lot more organized.
2792 * A new PartOf= unit dependency type has been introduced that
2793 may be used to group services in a natural way.
2795 * "systemctl enable" may now be used to enable instances of
2798 * journalctl now prints error log levels in red, and
2799 warning/notice log levels in bright white. It also supports
2800 filtering by log level now.
2802 * cgtop gained a new -n switch (similar to top), to configure
2803 the maximum number of iterations to run for. It also gained
2804 -b, to run in batch mode (accepting no input).
2806 * The suffix ".service" may now be omitted on most systemctl
2807 command lines involving service unit names.
2809 * There's a new bus call in logind to lock all sessions, as
2810 well as a loginctl verb for it "lock-sessions".
2812 * libsystemd-logind.so gained a new call sd_journal_perror()
2813 that works similar to libc perror() but logs to the journal
2814 and encodes structured information about the error number.
2816 * /etc/crypttab entries now understand the new keyfile-size=
2819 * shutdown(8) now can send a (configurable) wall message when
2820 a shutdown is cancelled.
2822 * The mount propagation mode for the root file system will now
2823 default to "shared", which is useful to make containers work
2824 nicely out-of-the-box so that they receive new mounts from
2825 the host. This can be undone locally by running "mount
2826 --make-rprivate /" if needed.
2828 * The prefdm.service file has been removed. Distributions
2829 should maintain this unit downstream if they intend to keep
2830 it around. However, we recommend writing normal unit files
2831 for display managers instead.
2833 * Since systemd is a crucial part of the OS we will now
2834 default to a number of compiler switches that improve
2835 security (hardening) such as read-only relocations, stack
2836 protection, and suchlike.
2838 * The TimeoutSec= setting for services is now split into
2839 TimeoutStartSec= and TimeoutStopSec= to allow configuration
2840 of individual time outs for the start and the stop phase of
2843 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
2844 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
2845 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
2846 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
2847 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
2848 Gundersen, Zbigniew Jędrzejewski-Szmek
2852 * The journal and id128 C APIs are now fully documented as man
2855 * Extra safety checks have been added when transitioning from
2856 the initial RAM disk to the main system to avoid accidental
2859 * /etc/crypttab entries now understand the new keyfile-offset=
2862 * systemctl -t can now be used to filter by unit load state.
2864 * The journal C API gained the new sd_journal_wait() call to
2865 make writing synchronous journal clients easier.
2867 * journalctl gained the new -D switch to show journals from a
2870 * journalctl now displays a special marker between log
2871 messages of two different boots.
2873 * The journal is now explicitly flushed to /var via a service
2874 systemd-journal-flush.service, rather than implicitly simply
2875 by seeing /var/log/journal to be writable.
2877 * journalctl (and the journal C APIs) can now match for much
2878 more complex expressions, with alternatives and
2881 * When transitioning from the initial RAM disk to the main
2882 system we will now kill all processes in a killing spree to
2883 ensure no processes stay around by accident.
2885 * Three new specifiers may be used in unit files: %u, %h, %s
2886 resolve to the user name, user home directory resp. user
2887 shell. This is useful for running systemd user instances.
2889 * We now automatically rotate journal files if their data
2890 object hash table gets a fill level > 75%. We also size the
2891 hash table based on the configured maximum file size. This
2892 together should lower hash collisions drastically and thus
2893 speed things up a bit.
2895 * journalctl gained the new "--header" switch to introspect
2896 header data of journal files.
2898 * A new setting SystemCallFilters= has been added to services
2899 which may be used to apply blacklists or whitelists to
2900 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
2902 * nspawn gained a new --link-journal= switch (and quicker: -j)
2903 to link the container journal with the host. This makes it
2904 very easy to centralize log viewing on the host for all
2905 guests while still keeping the journal files separated.
2907 * Many bugfixes and optimizations
2909 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
2910 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
2911 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
2916 * Several tools now understand kernel command line arguments,
2917 which are only read when run in an initial RAM disk. They
2918 usually follow closely their normal counterparts, but are
2921 * There's a new tool to analyze the readahead files that are
2922 automatically generated at boot. Use:
2924 /usr/lib/systemd/systemd-readahead analyze /.readahead
2926 * We now provide an early debug shell on tty9 if this enabled. Use:
2928 systemctl enable debug-shell.service
2930 * All plymouth related units have been moved into the Plymouth
2931 package. Please make sure to upgrade your Plymouth version
2934 * systemd-tmpfiles now supports getting passed the basename of
2935 a configuration file only, in which case it will look for it
2936 in all appropriate directories automatically.
2938 * udevadm info now takes a /dev or /sys path as argument, and
2939 does the right thing. Example:
2941 udevadm info /dev/sda
2942 udevadm info /sys/class/block/sda
2944 * systemctl now prints a warning if a unit is stopped but a
2945 unit that might trigger it continues to run. Example: a
2946 service is stopped but the socket that activates it is left
2949 * "systemctl status" will now mention if the log output was
2950 shortened due to rotation since a service has been started.
2952 * The journal API now exposes functions to determine the
2953 "cutoff" times due to rotation.
2955 * journald now understands SIGUSR1 and SIGUSR2 for triggering
2956 immediately flushing of runtime logs to /var if possible,
2957 resp. for triggering immediate rotation of the journal
2960 * It is now considered an error if a service is attempted to
2961 be stopped that is not loaded.
2963 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
2965 * systemd-analyze now supports Python 3
2967 * tmpfiles now supports cleaning up directories via aging
2968 where the first level dirs are always kept around but
2969 directories beneath it automatically aged. This is enabled
2970 by prefixing the age field with '~'.
2972 * Seat objects now expose CanGraphical, CanTTY properties
2973 which is required to deal with very fast bootups where the
2974 display manager might be running before the graphics drivers
2975 completed initialization.
2977 * Seat objects now expose a State property.
2979 * We now include RPM macros for service enabling/disabling
2980 based on the preset logic. We recommend RPM based
2981 distributions to make use of these macros if possible. This
2982 makes it simpler to reuse RPM spec files across
2985 * We now make sure that the collected systemd unit name is
2986 always valid when services log to the journal via
2989 * There's a new man page kernel-command-line(7) detailing all
2990 command line options we understand.
2992 * The fstab generator may now be disabled at boot by passing
2993 fstab=0 on the kernel command line.
2995 * A new kernel command line option modules-load= is now understood
2996 to load a specific kernel module statically, early at boot.
2998 * Unit names specified on the systemctl command line are now
2999 automatically escaped as needed. Also, if file system or
3000 device paths are specified they are automatically turned
3001 into the appropriate mount or device unit names. Example:
3003 systemctl status /home
3004 systemctl status /dev/sda
3006 * The SysVConsole= configuration option has been removed from
3007 system.conf parsing.
3009 * The SysV search path is no longer exported on the D-Bus
3012 * The Names= option is been removed from unit file parsing.
3014 * There's a new man page bootup(7) detailing the boot process.
3016 * Every unit and every generator we ship with systemd now
3017 comes with full documentation. The self-explanatory boot is
3020 * A couple of services gained "systemd-" prefixes in their
3021 name if they wrap systemd code, rather than only external
3022 code. Among them fsck@.service which is now
3023 systemd-fsck@.service.
3025 * The HaveWatchdog property has been removed from the D-Bus
3028 * systemd.confirm_spawn= on the kernel command line should now
3031 * There's a new man page crypttab(5) which details all options
3032 we actually understand.
3034 * systemd-nspawn gained a new --capability= switch to pass
3035 additional capabilities to the container.
3037 * timedated will now read known NTP implementation unit names
3038 from /usr/lib/systemd/ntp-units.d/*.list,
3039 systemd-timedated-ntp.target has been removed.
3041 * journalctl gained a new switch "-b" that lists log data of
3042 the current boot only.
3044 * The notify socket is in the abstract namespace again, in
3045 order to support daemons which chroot() at start-up.
3047 * There is a new Storage= configuration option for journald
3048 which allows configuration of where log data should go. This
3049 also provides a way to disable journal logging entirely, so
3050 that data collected is only forwarded to the console, the
3051 kernel log buffer or another syslog implementation.
3053 * Many bugfixes and optimizations
3055 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
3056 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
3057 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
3058 Shawn Landden, Tom Gundersen
3062 * "systemctl help <unit>" now shows the man page if one is
3065 * Several new man pages have been added.
3067 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
3068 MaxLevelConsole= can now be specified in
3069 journald.conf. These options allow reducing the amount of
3070 data stored on disk or forwarded by the log level.
3072 * TimerSlackNSec= can now be specified in system.conf for
3073 PID1. This allows system-wide power savings.
3075 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
3076 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
3081 * logind is now capable of (optionally) handling power and
3082 sleep keys as well as the lid switch.
3084 * journalctl now understands the syntax "journalctl
3085 /usr/bin/avahi-daemon" to get all log output of a specific
3088 * CapabilityBoundingSet= in system.conf now also influences
3089 the capability bound set of usermode helpers of the kernel.
3091 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
3092 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
3093 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
3094 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
3098 * Note that we skipped 139 releases here in order to set the
3099 new version to something that is greater than both udev's
3100 and systemd's most recent version number.
3102 * udev: all udev sources are merged into the systemd source tree now.
3103 All future udev development will happen in the systemd tree. It
3104 is still fully supported to use the udev daemon and tools without
3105 systemd running, like in initramfs or other init systems. Building
3106 udev though, will require the *build* of the systemd tree, but
3107 udev can be properly *run* without systemd.
3109 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
3110 should be used to create dead device nodes as workarounds for broken
3113 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
3114 no longer supported. udev_monitor_new_from_netlink() needs to be
3115 used to subscribe to events.
3117 * udev: when udevd is started by systemd, processes which are left
3118 behind by forking them off of udev rules, are unconditionally cleaned
3119 up and killed now after the event handling has finished. Services or
3120 daemons must be started as systemd services. Services can be
3121 pulled-in by udev to get started, but they can no longer be directly
3122 forked by udev rules.
3124 * udev: the daemon binary is called systemd-udevd now and installed
3125 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
3126 to adapt to that, create symlink, or rename the binary after building
3129 * libudev no longer provides these symbols:
3130 udev_monitor_from_socket()
3131 udev_queue_get_failed_list_entry()
3132 udev_get_{dev,sys,run}_path()
3133 The versions number was bumped and symbol versioning introduced.
3135 * systemd-loginctl and systemd-journalctl have been renamed
3136 to loginctl and journalctl to match systemctl.
3138 * The config files: /etc/systemd/systemd-logind.conf and
3139 /etc/systemd/systemd-journald.conf have been renamed to
3140 logind.conf and journald.conf. Package updates should rename
3141 the files to the new names on upgrade.
3143 * For almost all files the license is now LGPL2.1+, changed
3144 from the previous GPL2.0+. Exceptions are some minor stuff
3145 of udev (which will be changed to LGPL2.1 eventually, too),
3146 and the MIT licensed sd-daemon.[ch] library that is suitable
3147 to be used as drop-in files.
3149 * systemd and logind now handle system sleep states, in
3150 particular suspending and hibernating.
3152 * logind now implements a sleep/shutdown/idle inhibiting logic
3153 suitable for a variety of uses. Soonishly Lennart will blog
3154 about this in more detail.
3156 * var-run.mount and var-lock.mount are no longer provided
3157 (which prevously bind mounted these directories to their new
3158 places). Distributions which have not converted these
3159 directories to symlinks should consider stealing these files
3160 from git history and add them downstream.
3162 * We introduced the Documentation= field for units and added
3163 this to all our shipped units. This is useful to make it
3164 easier to explore the boot and the purpose of the various
3167 * All smaller setup units (such as
3168 systemd-vconsole-setup.service) now detect properly if they
3169 are run in a container and are skipped when
3170 appropriate. This guarantees an entirely noise-free boot in
3171 Linux container environments such as systemd-nspawn.
3173 * A framework for implementing offline system updates is now
3174 integrated, for details see:
3175 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
3177 * A new service type Type=idle is available now which helps us
3178 avoiding ugly interleaving of getty output and boot status
3181 * There's now a system-wide CapabilityBoundingSet= option to
3182 globally reduce the set of capabilities for the
3183 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
3184 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
3185 even CAP_NET_ADMIN system-wide for secure systems.
3187 * There are now system-wide DefaultLimitXXX= options to
3188 globally change the defaults of the various resource limits
3189 for all units started by PID 1.
3191 * Harald Hoyer's systemd test suite has been integrated into
3192 systemd which allows easy testing of systemd builds in qemu
3193 and nspawn. (This is really awesome! Ask us for details!)
3195 * The fstab parser is now implemented as generator, not inside
3198 * systemctl will now warn you if .mount units generated from
3199 /etc/fstab are out of date due to changes in fstab that
3200 have not been read by systemd yet.
3202 * systemd is now suitable for usage in initrds. Dracut has
3203 already been updated to make use of this. With this in place
3204 initrds get a slight bit faster but primarily are much
3205 easier to introspect and debug since "systemctl status" in
3206 the host system can be used to introspect initrd services,
3207 and the journal from the initrd is kept around too.
3209 * systemd-delta has been added, a tool to explore differences
3210 between user/admin configuration and vendor defaults.
3212 * PrivateTmp= now affects both /tmp and /var/tmp.
3214 * Boot time status messages are now much prettier and feature
3215 proper english language. Booting up systemd has never been
3218 * Read-ahead pack files now include the inode number of all
3219 files to pre-cache. When the inode changes the pre-caching
3220 is not attempted. This should be nicer to deal with updated
3221 packages which might result in changes of read-ahead
3224 * We now temporaritly lower the kernel's read_ahead_kb variable
3225 when collecting read-ahead data to ensure the kernel's
3226 built-in read-ahead does not add noise to our measurements
3227 of necessary blocks to pre-cache.
3229 * There's now RequiresMountsFor= to add automatic dependencies
3230 for all mounts necessary for a specific file system path.
3232 * MountAuto= and SwapAuto= have been removed from
3233 system.conf. Mounting file systems at boot has to take place
3236 * nspawn now learned a new switch --uuid= to set the machine
3237 ID on the command line.
3239 * nspawn now learned the -b switch to automatically search
3242 * vt102 is now the default TERM for serial TTYs, upgraded from
3245 * systemd-logind now works on VT-less systems.
3247 * The build tree has been reorganized. The individual
3248 components now have directories of their own.
3250 * A new condition type ConditionPathIsReadWrite= is now available.
3252 * nspawn learned the new -C switch to create cgroups for the
3253 container in other hierarchies.
3255 * We now have support for hardware watchdogs, configurable in
3258 * The scheduled shutdown logic now has a public API.
3260 * We now mount /tmp as tmpfs by default, but this can be
3261 masked and /etc/fstab can override it.
3263 * Since udisks does not make use of /media anymore we are not
3264 mounting a tmpfs on it anymore.
3266 * journalctl gained a new --local switch to only interleave
3267 locally generated journal files.
3269 * We can now load the IMA policy at boot automatically.
3271 * The GTK tools have been split off into a systemd-ui.
3273 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
3274 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
3275 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
3276 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
3277 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
3278 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
3279 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
3280 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
3281 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
3286 * This is mostly a bugfix release
3288 * Support optional initialization of the machine ID from the
3289 KVM or container configured UUID.
3291 * Support immediate reboots with "systemctl reboot -ff"
3293 * Show /etc/os-release data in systemd-analyze output
3295 * Many bugfixes for the journal, including endianness fixes and
3296 ensuring that disk space enforcement works
3298 * sd-login.h is C++ comptaible again
3300 * Extend the /etc/os-release format on request of the Debian
3303 * We now refuse non-UTF8 strings used in various configuration
3304 and unit files. This is done to ensure we do not pass invalid
3305 data over D-Bus or expose it elsewhere.
3307 * Register Mimo USB Screens as suitable for automatic seat
3310 * Read SELinux client context from journal clients in a race
3313 * Reorder configuration file lookup order. /etc now always
3314 overrides /run in order to allow the administrator to always
3315 and unconditionally override vendor supplied or
3316 automatically generated data.
3318 * The various user visible bits of the journal now have man
3319 pages. We still lack man pages for the journal API calls
3322 * We now ship all man pages in HTML format again in the
3325 Contributions from: Dave Reisner, Dirk Eibach, Frederic
3326 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
3327 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
3332 * This is mostly a bugfix release
3334 * systems lacking /etc/os-release are no longer supported.
3336 * Various functionality updates to libsystemd-login.so
3338 * Track class of PAM logins to distuingish greeters from
3341 Contributions from: Kay Sievers, Lennart Poettering, Michael
3346 * This is an important bugfix release for v41.
3348 * Building man pages is now optional which should be useful
3349 for those building systemd from git but unwilling to install
3352 * Watchdog support for supervising services is now usable. In
3353 a future release support for hardware watchdogs
3354 (i.e. /dev/watchdog) will be added building on this.
3356 * Service start rate limiting is now configurable and can be
3357 turned off per service. When a start rate limit is hit a
3358 reboot can automatically be triggered.
3360 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
3362 Contributions from: Benjamin Franzke, Bill Nottingham,
3363 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
3364 Schmidt, Michał Górny, Piotr Drąg
3368 * The systemd binary is installed /usr/lib/systemd/systemd now;
3369 An existing /sbin/init symlink needs to be adapted with the
3372 * The code that loads kernel modules has been ported to invoke
3373 libkmod directly, instead of modprobe. This means we do not
3374 support systems with module-init-tools anymore.
3376 * Watchdog support is now already useful, but still not
3379 * A new kernel command line option systemd.setenv= is
3380 understood to set system wide environment variables
3381 dynamically at boot.
3383 * We now limit the set of capabilities of systemd-journald.
3385 * We now set SIGPIPE to ignore by default, since it only is
3386 useful in shell pipelines, and has little use in general
3387 code. This can be disabled with IgnoreSIPIPE=no in unit
3390 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
3391 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
3396 * This is mostly a bugfix release
3398 * We now expose the reason why a service failed in the
3399 "Result" D-Bus property.
3401 * Rudimentary service watchdog support (will be completed over
3402 the next few releases.)
3404 * When systemd forks off in order execute some service we will
3405 now immediately changes its argv[0] to reflect which process
3406 it will execute. This is useful to minimize the time window
3407 with a generic argv[0], which makes bootcharts more useful
3409 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
3410 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
3411 Mike Kazantsev, Ray Strode
3415 * This is mostly a test release, but incorporates many
3418 * New systemd-cgtop tool to show control groups by their
3421 * Linking against libacl for ACLs is optional again. If
3422 disabled, support tracking device access for active logins
3423 goes becomes unavailable, and so does access to the user
3424 journals by the respective users.
3426 * If a group "adm" exists, journal files are automatically
3427 owned by them, thus allow members of this group full access
3428 to the system journal as well as all user journals.
3430 * The journal now stores the SELinux context of the logging
3431 client for all entries.
3433 * Add C++ inclusion guards to all public headers
3435 * New output mode "cat" in the journal to print only text
3436 messages, without any meta data like date or time.
3438 * Include tiny X server wrapper as a temporary stop-gap to
3439 teach XOrg udev display enumeration. This is used by display
3440 managers such as gdm, and will go away as soon as XOrg
3441 learned native udev hotplugging for display devices.
3443 * Add new systemd-cat tool for executing arbitrary programs
3444 with STDERR/STDOUT connected to the journal. Can also act as
3445 BSD logger replacement, and does so by default.
3447 * Optionally store all locally generated coredumps in the
3448 journal along with meta data.
3450 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
3451 writing short strings to files (for usage for /sys), and for
3452 creating symlinks, character and block device nodes.
3454 * New unit file option ControlGroupPersistent= to make cgroups
3455 persistent, following the mechanisms outlined in
3456 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
3458 * Support multiple local RTCs in a sane way
3460 * No longer monopolize IO when replaying readahead data on
3461 rotating disks, since we might starve non-file-system IO to
3462 death, since fanotify() will not see accesses done by blkid,
3465 * Do not show kernel threads in systemd-cgls anymore, unless
3466 requested with new -k switch.
3468 Contributions from: Dan Horák, Kay Sievers, Lennart
3469 Poettering, Michal Schmidt
3473 * This is mostly a test release, but incorporates many
3476 * The git repository moved to:
3477 git://anongit.freedesktop.org/systemd/systemd
3478 ssh://git.freedesktop.org/git/systemd/systemd
3480 * First release with the journal
3481 http://0pointer.de/blog/projects/the-journal.html
3483 * The journal replaces both systemd-kmsg-syslogd and
3484 systemd-stdout-bridge.
3486 * New sd_pid_get_unit() API call in libsystemd-logind
3488 * Many systemadm clean-ups
3490 * Introduce remote-fs-pre.target which is ordered before all
3491 remote mounts and may be used to start services before all
3494 * Added Mageia support
3496 * Add bash completion for systemd-loginctl
3498 * Actively monitor PID file creation for daemons which exit in
3499 the parent process before having finished writing the PID
3500 file in the daemon process. Daemons which do this need to be
3501 fixed (i.e. PID file creation must have finished before the
3502 parent exits), but we now react a bit more gracefully to them.
3504 * Add colourful boot output, mimicking the well-known output
3505 of existing distributions.
3507 * New option PassCredentials= for socket units, for
3508 compatibility with a recent kernel ABI breakage.
3510 * /etc/rc.local is now hooked in via a generator binary, and
3511 thus will no longer act as synchronization point during
3514 * systemctl list-unit-files now supports --root=.
3516 * systemd-tmpfiles now understands two new commands: z, Z for
3517 relabelling files according to the SELinux database. This is
3518 useful to apply SELinux labels to specific files in /sys,
3521 * Output of SysV services is now forwarded to both the console
3522 and the journal by default, not only just the console.
3524 * New man pages for all APIs from libsystemd-login.
3526 * The build tree got reorganized and a the build system is a
3527 lot more modular allowing embedded setups to specifically
3528 select the components of systemd they are interested in.
3530 * Support for Linux systems lacking the kernel VT subsystem is
3533 * configure's --with-rootdir= got renamed to
3534 --with-rootprefix= to follow the naming used by udev and
3537 * Unless specified otherwise we will now install to /usr instead
3538 of /usr/local by default.
3540 * Processes with '@' in argv[0][0] are now excluded from the
3541 final shut-down killing spree, following the logic explained
3543 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
3545 * All processes remaining in a service cgroup when we enter
3546 the START or START_PRE states are now killed with
3547 SIGKILL. That means it is no longer possible to spawn
3548 background processes from ExecStart= lines (which was never
3549 supported anyway, and bad style).
3551 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
3552 reloading of units together.
3554 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
3555 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
3556 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
3557 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
3558 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek