1 systemd System and Service Manager
5 * A new tool systemd-sysusers has been added. This tool
6 creates system users and groups in /etc/passwd and
7 /etc/group, based on static declarative system user/group
8 definitions in /usr/lib/sysusers.d/. This is useful to
9 enable factory resets and volatile systems that boot up with
10 an empty /etc directory, and thus need system users and
11 groups created during early boot. systemd now also ships
12 with two default sysusers.d/ files for the most basic
13 users and groups systemd and the core operating system
16 * A new tmpfiles snippet has been added that rebuilds the
17 essential files in /etc on boot, should they be missing.
19 * A directive for ensuring automatic clean-up of
20 /var/cache/man/ has been removed from the default
21 configuration. This line should now be shipped by the man
22 implementation. The necessary change has been made to the
23 man-db implementation. Note that you need to update your man
24 implementation to one that ships this line, otherwise no
25 automatic clean-up of /var/cache/man will take place.
27 * A new condition ConditionNeedsUpdate= has been added that
28 may conditionalize services to only run when /etc or /var
29 are "older" than the vendor operating system resources in
30 /usr. This is useful for reconstructing or updating /etc
31 after an offline update of /usr or a factory reset, on the
32 next reboot. Services that want to run once after such an
33 update or reset should use this condition and order
34 themselves before the new systemd-update-done.service, which
35 will mark the two directories as fully updated. A number of
36 service files have been added making use of this, to rebuild
37 the udev hardware database, the journald message catalog and
38 dynamic loader cache (ldconfig). The systemd-sysusers tool
39 described above also makes use of this now. With this in
40 place it is now possible to start up a minimal operating
41 system with /etc empty cleanly. Fore more information on the
42 concepts involved see this recent blog story:
44 http://0pointer.de/blog/projects/stateless.html
46 * A new system group "input" has been introduced, and all
47 input device nodes get this group assigned. This is useful
48 for system-level software to get access to input devices. It
49 complements what is already done for "audio" and "video".
51 * systemd-networkd learnt minimal DHCPv4 server support in
52 addition to the existing DHCPv4 client support. It also
53 learnt DHCPv6 client and IPv6 Router Solicitation client
54 support. The DHCPv4 client gained support for static routes
55 passed in from the server. Note that the [DHCPv4] section
56 known in older systemd-networkd versions has been renamed to
57 [DHCP] and is now also used by the DHCPv6 client. Existing
58 .network files using settings of this section need to be
61 * networkd gained support for vxlan virtual networks.
63 * networkd gained support for automatic allocation of address
64 ranges for interfaces from a system-wide pool of
65 addresses. This is useful for dynamically managing a large
66 number of interfaces with a single network configuration
67 file. In particular this is useful to easily assign
68 appropriate IP addresses to the veth links of a large number
71 * RPM macros for processing sysusers, sysctl and binfmt
72 drop-in snippets at package installation time have been
75 * The /etc/os-release file should now be placed in
76 /usr/lib/os-release. The old location is automatically
77 created as symlink. /usr/lib is the more appropriate
78 location of this file, since it shall actually describe the
79 vendor operating system shipped in /usr, and not the
80 configuration stored in /etc.
82 * .mount units gained a new boolean SloppyOptions= setting
83 that maps to mount(8)'s -s option which enables permissive
84 parsing of unknown mount options.
86 * tmpfiles learnt a new "L+" directive which creates a symlink
87 but (unlike "L") deletes a pre-existing file first, should
88 it already exist and not already be the correct
89 symlink. Similar, "b+", "c+" and "p+" directives have been
90 added as well, which create block and character devices, as
91 well as fifos in the filesystem, possibly removing any
92 pre-existing files of different types.
94 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
95 'argument' field (which so far specified the source to
96 symlink/copy the files from) is now optional. If ommited the
97 same file os copied from /usr/share/factory/ suffixed by the
98 full destination path. This is useful for populating /etc
99 with essential files, by copying them from vendor defaults
100 shipped in /usr/share/factory/etc.
102 * A new command "systemctl preset-all" has been added that
103 applies the service preset settings to all installed unit
104 files. A new switch --preset-mode= has been added that
105 controls whether only enable or only disable operations
108 * A new command "systemctl is-system-running" has been added
109 that allows checking the overall state of the system, for
110 example whether it is fully up an running.
112 * When the system boots up with an empty /etc, the equivalent
113 to "systemctl preset-all" is executed during early boot, to
114 make sure all default services are enabled after a factory
117 * systemd now contains a minimal preset file that enables the
118 most basic services systemd ships by default.
120 * Unit files' [Install] section gained a new DefaultInstance=
121 field for defining the default instance to create if a
122 template unit is enabled with no instance specified.
124 * A new passive target cryptsetup-pre.target has been added
125 that may be used by services that need to make they run and
126 finish before the first LUKS cryptographic device is set up.
128 * The /dev/loop-control and /dev/btrfs-control device nodes
129 are now owned by the "disk" group by default, opening up
130 access to this group.
132 * systemd-coredump will now automatically generate a
133 stack trace of all core dumps taking place on the system,
134 based on elfutils' libdw library. This stack trace is logged
137 * systemd-coredump may now optionally store coredumps directly
138 on disk (in /var/lib/systemd/coredump, possibly compressed),
139 instead of storing them unconditionally in the journal. This
140 mode is the new default. A new configuration file
141 /etc/systemd/coredump.conf has been added to configure this
142 and other parameters of systemd-coredump.
144 * coredumpctl gained a new "info" verb to show details about a
145 specific coredump. A new switch "-1" has also been added
146 that makes sure to only show information about the most
147 recent entry instead of all entries. Also, as the tool is
148 generally useful now the "systemd-" prefix of the binary
149 name has been removed. Distributions that want to maintain
150 compatibility with the old name should add a symlink from
151 the old name to the new name.
153 * journald's SplitMode= now defaults to "uid". This makes sure
154 that unpriviliged users can access their own coredumps with
155 coredumpctl without restrictions.
157 * New kernel command line options "systemd.wants=" (for
158 pulling an additional unit during boot), "systemd.mask="
159 (for masking a specific unit for the boot), and
160 "systemd.debug-shell" (for enabling the debug shell on tty9)
161 have been added. This is implemented in the new generator
162 "systemd-debug-generator".
164 * systemd-nspawn will now by default filter a couple of
165 syscalls for containers, among them those required for
166 kernel module loading, direct x86 IO port access, swap
167 management, and kexec. Most importantly though
168 open_by_handle_at() is now prohibited for containers,
169 closing a hole similar to a recently discussed vulnerability
170 in docker regarding access to files on file hierarchies the
171 container should normally not have access to. Note that for
172 nspawn we generally make no security claims anyway (and
173 this is explicitly documented in the man page), so this is
174 just a fix for one of the most obvious problems.
176 * A new man page file-hierarchy(7) has been added that
177 contains a minimized, modernized version of the file system
178 layout systemd expects, similar in style to the FHS
179 specification or hier(5).
181 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
182 longer done. Since the directory now has a per-user size
183 limit, and is cleaned on logout this appears unnecessary,
184 in particular since this now brings the lifecycle of this
185 directory closer in line with how IPC objects are handled.
187 * systemd.pc now exports a number of additional directories,
188 including $libdir (which is useful to identify the library
189 path for the primary architecture of the system), and a
190 couple of drop-in directories.
194 * As an experimental feature, udev now tries to lock the
195 disk device node (flock(LOCK_SH|LOCK_NB)) while it
196 executes events for the disk or any of its partitions.
197 Applications like partitioning programs can lock the
198 disk device node (flock(LOCK_EX)) and claim temporary
199 device ownership that way; udev will entirely skip all event
200 handling for this disk and its partitions. If the disk
201 was opened for writing, the close will trigger a partition
202 table rescan in udev's "watch" facility, and if needed
203 synthesize "change" events for the disk and all its partitions.
204 This is now unconditionally enabled, and if it turns out to
205 cause major problems, we might turn it on only for specific
206 devices, or might need to disable it entirely. Device Mapper
207 devices are excluded from this logic.
209 * We temporarily dropped the "-l" switch for fsck invocations,
210 since they collide with the flock() logic above. util-linux
211 upstream has been changed already to avoid this conflict,
212 and we will readd "-l" as soon as util-linux with this
213 change has been released.
215 * The dependency on libattr has been removed. Since a long
216 time, the extended attribute calls have moved to glibc, and
217 libattr is thus unnecessary.
219 * Virtualization detection works without priviliges now. This
220 means the systemd-detect-virt binary no longer requires
221 CAP_SYS_PTRACE file capabilities, and our daemons can run
222 with fewer privileges.
224 * systemd-networkd now runs under its own "systemd-network"
225 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
226 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
227 loses the ability to write to files owned by root this way.
229 * Similar, systemd-resolved now runs under its own
230 "systemd-resolve" user with no capabilities remaining.
232 * Similar, systemd-bus-proxyd now runs under its own
233 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
235 * systemd-networkd gained support for setting up "veth"
236 virtual ethernet devices for container connectivity, as well
237 as GRE and VTI tunnels.
239 * systemd-networkd will no longer automatically attempt to
240 manually load kernel modules necessary for certain tunnel
241 transports. Instead, it is assumed the kernel loads them
242 automatically when required. This only works correctly on
243 very new kernels. On older kernels, please consider adding
244 the kernel modules to /etc/modules-load.d/ as a work-around.
246 * The resolv.conf file systemd-resolved generates has been
247 moved to /run/systemd/resolve/. If you have a symlink from
248 /etc/resolv.conf, it might be necessary to correct it.
250 * Two new service settings, ProtectedHome= and ProtectedSystem=,
251 have been added. When enabled, they will make the user data
252 (such as /home) inaccessible or read-only and the system
253 (such as /usr) read-only, for specific services. This allows
254 very light-weight per-service sandboxing to avoid
255 modifications of user data or system files from
256 services. These two new switches have been enabled for all
257 of systemd's long-running services, where appropriate.
259 * Socket units gained new SocketUser= and SocketGroup=
260 settings to set the owner user and group of AF_UNIX sockets
261 and FIFOs in the file system.
263 * Socket units gained a new RemoveOnStop= setting. If enabled,
264 all FIFOS and sockets in the file system will be removed
265 when the specific socket unit is stopped.
267 * Socket units gained a new Symlinks= setting. It takes a list
268 of symlinks to create to file system sockets or FIFOs
269 created by the specific Unix sockets. This is useful to
270 manage symlinks to socket nodes with the same life-cycle as
273 * The /dev/log socket and /dev/initctl FIFO have been moved to
274 /run, and have been replaced by symlinks. This allows
275 connecting to these facilities even if PrivateDevices=yes is
276 used for a service (which makes /dev/log itself unavailable,
277 but /run is left). This also has the benefit of ensuring
278 that /dev only contains device nodes, directories and
279 symlinks, and nothing else.
281 * sd-daemon gained two new calls sd_pid_notify() and
282 sd_pid_notifyf(). They are similar to sd_notify() and
283 sd_notifyf(), but allow overriding of the source PID of
284 notification messages if permissions permit this. This is
285 useful to send notify messages on behalf of a different
286 process (for example, the parent process). The
287 systemd-notify tool has been updated to make use of this
288 when sending messages (so that notification messages now
289 originate from the shell script invoking systemd-notify and
290 not the systemd-notify process itself. This should minimize
291 a race where systemd fails to associate notification
292 messages to services when the originating process already
295 * A new "on-abnormal" setting for Restart= has been added. If
296 set, it will result in automatic restarts on all "abnormal"
297 reasons for a process to exit, which includes unclean
298 signals, core dumps, timeouts and watchdog timeouts, but
299 does not include clean and unclean exit codes or clean
300 signals. Restart=on-abnormal is an alternative for
301 Restart=on-failure for services that shall be able to
302 terminate and avoid restarts on certain errors, by
303 indicating so with an unclean exit code. Restart=on-failure
304 or Restart=on-abnormal is now the recommended setting for
305 all long-running services.
307 * If the InaccessibleDirectories= service setting points to a
308 mount point (or if there are any submounts contained within
309 it), it is now attempted to completely unmount it, to make
310 the file systems truly unavailable for the respective
313 * The ReadOnlyDirectories= service setting and
314 systemd-nspawn's --read-only parameter are now recursively
315 applied to all submounts, too.
317 * Mount units may now be created transiently via the bus APIs.
319 * The support for SysV and LSB init scripts has been removed
320 from the systemd daemon itself. Instead, it is now
321 implemented as a generator that creates native systemd units
322 from these scripts when needed. This enables us to remove a
323 substantial amount of legacy code from PID 1, following the
324 fact that many distributions only ship a very small number
325 of LSB/SysV init scripts nowadays.
327 * Priviliged Xen (dom0) domains are not considered
328 virtualization anymore by the virtualization detection
329 logic. After all, they generally have unrestricted access to
330 the hardware and usually are used to manage the unprivileged
333 * systemd-tmpfiles gained a new "C" line type, for copying
334 files or entire directories.
336 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
337 lines. So far, they have been non-globbing versions of the
338 latter, and have thus been redundant. In future, it is
339 recommended to only use "z". "m" has hence been removed
340 from the documentation, even though it stays supported.
342 * A tmpfiles snippet to recreate the most basic structure in
343 /var has been added. This is enough to create the /var/run →
344 /run symlink and create a couple of structural
345 directories. This allows systems to boot up with an empty or
346 volatile /var. Of course, while with this change, the core OS
347 now is capable with dealing with a volatile /var, not all
348 user services are ready for it. However, we hope that sooner
349 or later, many service daemons will be changed upstream so
350 that they are able to automatically create their necessary
351 directories in /var at boot, should they be missing. This is
352 the first step to allow state-less systems that only require
353 the vendor image for /usr to boot.
355 * systemd-nspawn has gained a new --tmpfs= switch to mount an
356 empty tmpfs instance to a specific directory. This is
357 particularly useful for making use of the automatic
358 reconstruction of /var (see above), by passing --tmpfs=/var.
360 * Access modes specified in tmpfiles snippets may now be
361 prefixed with "~", which indicates that they shall be masked
362 by whether the existing file or directly is currently
363 writable, readable or executable at all. Also, if specified,
364 the sgid/suid/sticky bits will be masked for all
367 * A new passive target unit "network-pre.target" has been
368 added which is useful for services that shall run before any
369 network is configured, for example firewall scripts.
371 * The "floppy" group that previously owned the /dev/fd*
372 devices is no longer used. The "disk" group is now used
373 instead. Distributions should probably deprecate usage of
376 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
377 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
378 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
379 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
380 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
381 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
384 -- Berlin, 2014-06-11
388 * A new "systemd-timesyncd" daemon has been added for
389 synchronizing the system clock across the network. It
390 implements an SNTP client. In contrast to NTP
391 implementations such as chrony or the NTP reference server,
392 this only implements a client side, and does not bother with
393 the full NTP complexity, focusing only on querying time from
394 one remote server and synchronizing the local clock to
395 it. Unless you intend to serve NTP to networked clients or
396 want to connect to local hardware clocks, this simple NTP
397 client should be more than appropriate for most
398 installations. The daemon runs with minimal privileges, and
399 has been hooked up with networkd to only operate when
400 network connectivity is available. The daemon saves the
401 current clock to disk every time a new NTP sync has been
402 acquired, and uses this to possibly correct the system clock
403 early at bootup, in order to accommodate for systems that
404 lack an RTC such as the Raspberry Pi and embedded devices,
405 and to make sure that time monotonically progresses on these
406 systems, even if it is not always correct. To make use of
407 this daemon, a new system user and group "systemd-timesync"
408 needs to be created on installation of systemd.
410 * The queue "seqnum" interface of libudev has been disabled, as
411 it was generally incompatible with device namespacing as
412 sequence numbers of devices go "missing" if the devices are
413 part of a different namespace.
415 * "systemctl list-timers" and "systemctl list-sockets" gained
416 a --recursive switch for showing units of these types also
417 for all local containers, similar in style to the already
418 supported --recursive switch for "systemctl list-units".
420 * A new RebootArgument= setting has been added for service
421 units, which may be used to specify a kernel reboot argument
422 to use when triggering reboots with StartLimitAction=.
424 * A new FailureAction= setting has been added for service
425 units which may be used to specify an operation to trigger
426 when a service fails. This works similarly to
427 StartLimitAction=, but unlike it, controls what is done
428 immediately rather than only after several attempts to
429 restart the service in question.
431 * hostnamed got updated to also expose the kernel name,
432 release, and version on the bus. This is useful for
433 executing commands like hostnamectl with the -H switch.
434 systemd-analyze makes use of this to properly display
435 details when running non-locally.
437 * The bootchart tool can now show cgroup information in the
440 * The CFS CPU quota cgroup attribute is now exposed for
441 services. The new CPUQuota= switch has been added for this
442 which takes a percentage value. Setting this will have the
443 result that a service may never get more CPU time than the
444 specified percentage, even if the machine is otherwise idle.
446 * systemd-networkd learned IPIP and SIT tunnel support.
448 * LSB init scripts exposing a dependency on $network will now
449 get a dependency on network-online.target rather than simply
450 network.target. This should bring LSB handling closer to
451 what it was on SysV systems.
453 * A new fsck.repair= kernel option has been added to control
454 how fsck shall deal with unclean file systems at boot.
456 * The (.ini) configuration file parser will now silently
457 ignore sections whose name begins with "X-". This may be
458 used to maintain application-specific extension sections in unit
461 * machined gained a new API to query the IP addresses of
462 registered containers. "machinectl status" has been updated
463 to show these addresses in its output.
465 * A new call sd_uid_get_display() has been added to the
466 sd-login APIs for querying the "primary" session of a
467 user. The "primary" session of the user is elected from the
468 user's sessions and generally a graphical session is
469 preferred over a text one.
471 * A minimal systemd-resolved daemon has been added. It
472 currently simply acts as a companion to systemd-networkd and
473 manages resolv.conf based on per-interface DNS
474 configuration, possibly supplied via DHCP. In the long run
475 we hope to extend this into a local DNSSEC enabled DNS and
478 * The systemd-networkd-wait-online tool is now enabled by
479 default. It will delay network-online.target until a network
480 connection has been configured. The tool primarily integrates
481 with networkd, but will also make a best effort to make sense
482 of network configuration performed in some other way.
484 * Two new service options StartupCPUShares= and
485 StartupBlockIOWeight= have been added that work similarly to
486 CPUShares= and BlockIOWeight= however only apply during
487 system startup. This is useful to prioritize certain services
488 differently during bootup than during normal runtime.
490 * hostnamed has been changed to prefer the statically
491 configured hostname in /etc/hostname (unless set to
492 'localhost' or empty) over any dynamic one supplied by
493 dhcp. With this change, the rules for picking the hostname
494 match more closely the rules of other configuration settings
495 where the local administrator's configuration in /etc always
496 overrides any other settings.
498 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
499 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
500 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
501 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
502 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
503 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
504 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
505 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
506 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
507 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
508 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
509 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
510 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
511 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
512 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
513 Lindskog, WaLyong Cho, Will Woods, Zbigniew
516 -- Beijing, 2014-05-28
520 * When restoring the screen brightness at boot, stay away from
521 the darkest setting or from the lowest 5% of the available
522 range, depending on which is the larger value of both. This
523 should effectively protect the user from rebooting into a
524 black screen, should the brightness have been set to minimum
527 * sd-login gained a new sd_machine_get_class() call to
528 determine the class ("vm" or "container") of a machine
529 registered with machined.
531 * sd-login gained new calls
532 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
533 to query the identity of the peer of a local AF_UNIX
534 connection. They operate similarly to their sd_pid_get_xyz()
537 * PID 1 will now maintain a system-wide system state engine
538 with the states "starting", "running", "degraded",
539 "maintenance", "stopping". These states are bound to system
540 startup, normal runtime, runtime with at least one failed
541 service, rescue/emergency mode and system shutdown. This
542 state is shown in the "systemctl status" output when no unit
543 name is passed. It is useful to determine system state, in
544 particularly when doing so for many systems or containers at
547 * A new command "list-machines" has been added to "systemctl"
548 that lists all local OS containers and shows their system
549 state (see above), if systemd runs inside of them.
551 * systemctl gained a new "-r" switch to recursively enumerate
552 units on all local containers, when used with the
553 "list-unit" command (which is the default one that is
554 executed when no parameters are specified).
556 * The GPT automatic partition discovery logic will now honour
557 two GPT partition flags: one may be set on a partition to
558 cause it to be mounted read-only, and the other may be set
559 on a partition to ignore it during automatic discovery.
561 * Two new GPT type UUIDs have been added for automatic root
562 partition discovery, for 32-bit and 64-bit ARM. This is not
563 particularly useful for discovering the root directory on
564 these architectures during bare-metal boots (since UEFI is
565 not common there), but still very useful to allow booting of
566 ARM disk images in nspawn with the -i option.
568 * MAC addresses of interfaces created with nspawn's
569 --network-interface= switch will now be generated from the
570 machine name, and thus be stable between multiple invocations
573 * logind will now automatically remove all IPC objects owned
574 by a user if she or he fully logs out. This makes sure that
575 users who are logged out cannot continue to consume IPC
576 resources. This covers SysV memory, semaphores and message
577 queues as well as POSIX shared memory and message
578 queues. Traditionally, SysV and POSIX IPC had no life-cycle
579 limits. With this functionality, that is corrected. This may
580 be turned off by using the RemoveIPC= switch of logind.conf.
582 * The systemd-machine-id-setup and tmpfiles tools gained a
583 --root= switch to operate on a specific root directory,
586 * journald can now forward logged messages to the TTYs of all
587 logged in users ("wall"). This is the default for all
588 emergency messages now.
590 * A new tool systemd-journal-remote has been added to stream
591 journal log messages across the network.
593 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
594 controller trees are mounted into it. Note that the
595 directories mounted beneath it are not read-only. This is a
596 security measure and is particularly useful because glibc
597 actually includes a search logic to pick any tmpfs it can
598 find to implement shm_open() if /dev/shm is not available
599 (which it might very well be in namespaced setups).
601 * machinectl gained a new "poweroff" command to cleanly power
602 down a local OS container.
604 * The PrivateDevices= unit file setting will now also drop the
605 CAP_MKNOD capability from the capability bound set, and
606 imply DevicePolicy=closed.
608 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
609 comprehensively on all long-running systemd services where
612 * systemd-udevd will now run in a disassociated mount
613 namespace. To mount directories from udev rules, make sure to
614 pull in mount units via SYSTEMD_WANTS properties.
616 * The kdbus support gained support for uploading policy into
617 the kernel. sd-bus gained support for creating "monitoring"
618 connections that can eavesdrop into all bus communication
619 for debugging purposes.
621 * Timestamps may now be specified in seconds since the UNIX
622 epoch Jan 1st, 1970 by specifying "@" followed by the value
625 * Native tcpwrap support in systemd has been removed. tcpwrap
626 is old code, not really maintained anymore and has serious
627 shortcomings, and better options such as firewalls
628 exist. For setups that require tcpwrap usage, please
629 consider invoking your socket-activated service via tcpd,
630 like on traditional inetd.
632 * A new system.conf configuration option
633 DefaultTimerAccuracySec= has been added that controls the
634 default AccuracySec= setting of .timer units.
636 * Timer units gained a new WakeSystem= switch. If enabled,
637 timers configured this way will cause the system to resume
638 from system suspend (if the system supports that, which most
641 * Timer units gained a new Persistent= switch. If enabled,
642 timers configured this way will save to disk when they have
643 been last triggered. This information is then used on next
644 reboot to possible execute overdue timer events, that
645 could not take place because the system was powered off.
646 This enables simple anacron-like behaviour for timer units.
648 * systemctl's "list-timers" will now also list the time a
649 timer unit was last triggered in addition to the next time
650 it will be triggered.
652 * systemd-networkd will now assign predictable IPv4LL
653 addresses to its local interfaces.
655 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
656 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
657 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
658 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
659 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
660 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
661 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
662 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
665 -- Berlin, 2014-03-25
669 * A new unit file setting RestrictAddressFamilies= has been
670 added to restrict which socket address families unit
671 processes gain access to. This takes address family names
672 like "AF_INET" or "AF_UNIX", and is useful to minimize the
673 attack surface of services via exotic protocol stacks. This
674 is built on seccomp system call filters.
676 * Two new unit file settings RuntimeDirectory= and
677 RuntimeDirectoryMode= have been added that may be used to
678 manage a per-daemon runtime directories below /run. This is
679 an alternative for setting up directory permissions with
680 tmpfiles snippets, and has the advantage that the runtime
681 directory's lifetime is bound to the daemon runtime and that
682 the daemon starts up with an empty directory each time. This
683 is particularly useful when writing services that drop
684 privileges using the User= or Group= setting.
686 * The DeviceAllow= unit setting now supports globbing for
687 matching against device group names.
689 * The systemd configuration file system.conf gained new
690 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
691 DefaultMemoryAccounting= to globally turn on/off accounting
692 for specific resources (cgroups) for all units. These
693 settings may still be overridden individually in each unit
696 * systemd-gpt-auto-generator is now able to discover /srv and
697 root partitions in addition to /home and swap partitions. It
698 also supports LUKS-encrypted partitions now. With this in
699 place, automatic discovery of partitions to mount following
700 the Discoverable Partitions Specification
701 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
702 is now a lot more complete. This allows booting without
703 /etc/fstab and without root= on the kernel command line on
704 systems prepared appropriately.
706 * systemd-nspawn gained a new --image= switch which allows
707 booting up disk images and Linux installations on any block
708 device that follow the Discoverable Partitions Specification
709 (see above). This means that installations made with
710 appropriately updated installers may now be started and
711 deployed using container managers, completely
712 unmodified. (We hope that libvirt-lxc will add support for
713 this feature soon, too.)
715 * systemd-nspawn gained a new --network-macvlan= setting to
716 set up a private macvlan interface for the
717 container. Similarly, systemd-networkd gained a new
718 Kind=macvlan setting in .netdev files.
720 * systemd-networkd now supports configuring local addresses
723 * A new tool systemd-network-wait-online has been added to
724 synchronously wait for network connectivity using
727 * The sd-bus.h bus API gained a new sd_bus_track object for
728 tracking the life-cycle of bus peers. Note that sd-bus.h is
729 still not a public API though (unless you specify
730 --enable-kdbus on the configure command line, which however
731 voids your warranty and you get no API stability guarantee).
733 * The $XDG_RUNTIME_DIR runtime directories for each user are
734 now individual tmpfs instances, which has the benefit of
735 introducing separate pools for each user, with individual
736 size limits, and thus making sure that unprivileged clients
737 can no longer negatively impact the system or other users by
738 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
739 RuntimeDirectorySize= has been introduced that allows
740 controlling the default size limit for all users. It
741 defaults to 10% of the available physical memory. This is no
742 replacement for quotas on tmpfs though (which the kernel
743 still does not support), as /dev/shm and /tmp are still
744 shared resources used by both the system and unprivileged
747 * logind will now automatically turn off automatic suspending
748 on laptop lid close when more than one display is
749 connected. This was previously expected to be implemented
750 individually in desktop environments (such as GNOME),
751 however has been added to logind now, in order to fix a
752 boot-time race where a desktop environment might not have
753 been started yet and thus not been able to take an inhibitor
754 lock at the time where logind already suspends the system
757 * logind will now wait at least 30s after each system
758 suspend/resume cycle, and 3min after system boot before
759 suspending the system due to a closed laptop lid. This
760 should give USB docking stations and similar enough time to
761 be probed and configured after system resume and boot in
762 order to then act as suspend blocker.
764 * systemd-run gained a new --property= setting which allows
765 initialization of resource control properties (and others)
766 for the created scope or service unit. Example: "systemd-run
767 --property=BlockIOWeight=10 updatedb" may be used to run
768 updatedb at a low block IO scheduling weight.
770 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
771 now also work in --scope mode.
773 * When systemd is compiled with kdbus support, basic support
774 for enforced policies is now in place. (Note that enabling
775 kdbus still voids your warranty and no API compatibility
778 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
779 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
780 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
781 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
782 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
783 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
784 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
785 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
786 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
787 Zbigniew Jędrzejewski-Szmek
789 -- Berlin, 2014-03-12
793 * systemd will now relabel /dev after loading the SMACK policy
794 according to SMACK rules.
796 * A new unit file option AppArmorProfile= has been added to
797 set the AppArmor profile for the processes of a unit.
799 * A new condition check ConditionArchitecture= has been added
800 to conditionalize units based on the system architecture, as
801 reported by uname()'s "machine" field.
803 * systemd-networkd now supports matching on the system
804 virtualization, architecture, kernel command line, host name
807 * logind is now a lot more aggressive when suspending the
808 machine due to a closed laptop lid. Instead of acting only
809 on the lid close action, it will continuously watch the lid
810 status and act on it. This is useful for laptops where the
811 power button is on the outside of the chassis so that it can
812 be reached without opening the lid (such as the Lenovo
813 Yoga). On those machines, logind will now immediately
814 re-suspend the machine if the power button has been
815 accidentally pressed while the laptop was suspended and in a
818 * logind will now watch SW_DOCK switches and inhibit reaction
819 to the lid switch if it is pressed. This means that logind
820 will not suspend the machine anymore if the lid is closed
821 and the system is docked, if the laptop supports SW_DOCK
822 notifications via the input layer. Note that ACPI docking
823 stations do not generate this currently. Also note that this
824 logic is usually not fully sufficient and Desktop
825 Environments should take a lid switch inhibitor lock when an
826 external display is connected, as systemd will not watch
829 * nspawn will now make use of the devices cgroup controller by
830 default, and only permit creation of and access to the usual
831 API device nodes like /dev/null or /dev/random, as well as
832 access to (but not creation of) the pty devices.
834 * We will now ship a default .network file for
835 systemd-networkd that automatically configures DHCP for
836 network interfaces created by nspawn's --network-veth or
837 --network-bridge= switches.
839 * systemd will now understand the usual M, K, G, T suffixes
840 according to SI conventions (i.e. to the base 1000) when
841 referring to throughput and hardware metrics. It will stay
842 with IEC conventions (i.e. to the base 1024) for software
843 metrics, according to what is customary according to
844 Wikipedia. We explicitly document which base applies for
845 each configuration option.
847 * The DeviceAllow= setting in unit files now supports a syntax
848 to whitelist an entire group of devices node majors at once,
849 based on the /proc/devices listing. For example, with the
850 string "char-pts", it is now possible to whitelist all
851 current and future pseudo-TTYs at once.
853 * sd-event learned a new "post" event source. Event sources of
854 this type are triggered by the dispatching of any event
855 source of a type that is not "post". This is useful for
856 implementing clean-up and check event sources that are
857 triggered by other work being done in the program.
859 * systemd-networkd is no longer statically enabled, but uses
860 the usual [Install] sections so that it can be
861 enabled/disabled using systemctl. It still is enabled by
864 * When creating a veth interface pair with systemd-nspawn, the
865 host side will now be prefixed with "vb-" if
866 --network-bridge= is used, and with "ve-" if --network-veth
867 is used. This way, it is easy to distinguish these cases on
868 the host, for example to apply different configuration to
869 them with systemd-networkd.
871 * The compatibility libraries for libsystemd-journal.so,
872 libsystem-id128.so, libsystemd-login.so and
873 libsystemd-daemon.so do not make use of IFUNC
874 anymore. Instead, we now build libsystemd.so multiple times
875 under these alternative names. This means that the footprint
876 is drastically increased, but given that these are
877 transitional compatibility libraries, this should not matter
878 much. This change has been made necessary to support the ARM
879 platform for these compatibility libraries, as the ARM
880 toolchain is not really at the same level as the toolchain
881 for other architectures like x86 and does not support
882 IFUNC. Please make sure to use --enable-compat-libs only
883 during a transitional period!
885 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
886 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
887 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
888 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
889 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
890 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
891 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
892 Zbigniew Jędrzejewski-Szmek
894 -- Berlin, 2014-02-24
898 * A new component "systemd-networkd" has been added that can
899 be used to configure local network interfaces statically or
900 via DHCP. It is capable of bringing up bridges, VLANs, and
901 bonding. Currently, no hook-ups for interactive network
902 configuration are provided. Use this for your initrd,
903 container, embedded, or server setup if you need a simple,
904 yet powerful, network configuration solution. This
905 configuration subsystem is quite nifty, as it allows wildcard
906 hotplug matching in interfaces. For example, with a single
907 configuration snippet, you can configure that all Ethernet
908 interfaces showing up are automatically added to a bridge,
909 or similar. It supports link-sensing and more.
911 * A new tool "systemd-socket-proxyd" has been added which can
912 act as a bidirectional proxy for TCP sockets. This is
913 useful for adding socket activation support to services that
914 do not actually support socket activation, including virtual
915 machines and the like.
917 * Add a new tool to save/restore rfkill state on
920 * Save/restore state of keyboard backlights in addition to
921 display backlights on shutdown/boot.
923 * udev learned a new SECLABEL{} construct to label device
924 nodes with a specific security label when they appear. For
925 now, only SECLABEL{selinux} is supported, but the syntax is
926 prepared for additional security frameworks.
928 * udev gained a new scheme to configure link-level attributes
929 from files in /etc/systemd/network/*.link. These files can
930 match against MAC address, device path, driver name and type,
931 and will apply attributes like the naming policy, link speed,
932 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
933 address assignment policy (randomized, ...).
935 * The configuration of network interface naming rules for
936 "permanent interface names" has changed: a new NamePolicy=
937 setting in the [Link] section of .link files determines the
938 priority of possible naming schemes (onboard, slot, mac,
939 path). The default value of this setting is determined by
940 /usr/lib/net/links/99-default.link. Old
941 80-net-name-slot.rules udev configuration file has been
942 removed, so local configuration overriding this file should
943 be adapated to override 99-default.link instead.
945 * When the User= switch is used in a unit file, also
946 initialize $SHELL= based on the user database entry.
948 * systemd no longer depends on libdbus. All communication is
949 now done with sd-bus, systemd's low-level bus library
952 * kdbus support has been added to PID 1 itself. When kdbus is
953 enabled, this causes PID 1 to set up the system bus and
954 enable support for a new ".busname" unit type that
955 encapsulates bus name activation on kdbus. It works a little
956 bit like ".socket" units, except for bus names. A new
957 generator has been added that converts classic dbus1 service
958 activation files automatically into native systemd .busname
961 * sd-bus: add a light-weight vtable implementation that allows
962 defining objects on the bus with a simple static const
963 vtable array of its methods, signals and properties.
965 * systemd will not generate or install static dbus
966 introspection data anymore to /usr/share/dbus-1/interfaces,
967 as the precise format of these files is unclear, and
968 nothing makes use of it.
970 * A proxy daemon is now provided to proxy clients connecting
971 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
972 compatibility with classic D-Bus.
974 * A bus driver implementation has been added that supports the
975 classic D-Bus bus driver calls on kdbus, also for
976 compatibility purposes.
978 * A new API "sd-event.h" has been added that implements a
979 minimal event loop API built around epoll. It provides a
980 couple of features that direct epoll usage is lacking:
981 prioritization of events, scales to large numbers of timer
982 events, per-event timer slack (accuracy), system-wide
983 coalescing of timer events, exit handlers, watchdog
984 supervision support using systemd's sd_notify() API, child
987 * A new API "sd-rntl.h" has been added that provides an API
988 around the route netlink interface of the kernel, similar in
991 * A new API "sd-dhcp-client.h" has been added that provides a
992 small DHCPv4 client-side implementation. This is used by
995 * There is a new kernel command line option
996 "systemd.restore_state=0|1". When set to "0", none of the
997 systemd tools will restore saved runtime state to hardware
998 devices. More specifically, the rfkill and backlight states
1001 * The FsckPassNo= compatibility option in mount/service units
1002 has been removed. The fstab generator will now add the
1003 necessary dependencies automatically, and does not require
1004 PID1's support for that anymore.
1006 * journalctl gained a new switch, --list-boots, that lists
1007 recent boots with their times and boot IDs.
1009 * The various tools like systemctl, loginctl, timedatectl,
1010 busctl, systemd-run, ... have gained a new switch "-M" to
1011 connect to a specific, local OS container (as direct
1012 connection, without requiring SSH). This works on any
1013 container that is registered with machined, such as those
1014 created by libvirt-lxc or nspawn.
1016 * systemd-run and systemd-analyze also gained support for "-H"
1017 to connect to remote hosts via SSH. This is particularly
1018 useful for systemd-run because it enables queuing of jobs
1019 onto remote systems.
1021 * machinectl gained a new command "login" to open a getty
1022 login in any local container. This works with any container
1023 that is registered with machined (such as those created by
1024 libvirt-lxc or nspawn), and which runs systemd inside.
1026 * machinectl gained a new "reboot" command that may be used to
1027 trigger a reboot on a specific container that is registered
1028 with machined. This works on any container that runs an init
1029 system of some kind.
1031 * systemctl gained a new "list-timers" command to print a nice
1032 listing of installed timer units with the times they elapse
1035 * Alternative reboot() parameters may now be specified on the
1036 "systemctl reboot" command line and are passed to the
1037 reboot() system call.
1039 * systemctl gained a new --job-mode= switch to configure the
1040 mode to queue a job with. This is a more generic version of
1041 --fail, --irreversible, and --ignore-dependencies, which are
1042 still available but not advertised anymore.
1044 * /etc/systemd/system.conf gained new settings to configure
1045 various default timeouts of units, as well as the default
1046 start limit interval and burst. These may still be overridden
1049 * PID1 will now export on the bus profile data of the security
1050 policy upload process (such as the SELinux policy upload to
1053 * journald: when forwarding logs to the console, include
1054 timestamps (following the setting in
1055 /sys/module/printk/parameters/time).
1057 * OnCalendar= in timer units now understands the special
1058 strings "yearly" and "annually". (Both are equivalent)
1060 * The accuracy of timer units is now configurable with the new
1061 AccuracySec= setting. It defaults to 1min.
1063 * A new dependency type JoinsNamespaceOf= has been added that
1064 allows running two services within the same /tmp and network
1065 namespace, if PrivateNetwork= or PrivateTmp= are used.
1067 * A new command "cat" has been added to systemctl. It outputs
1068 the original unit file of a unit, and concatenates the
1069 contents of additional "drop-in" unit file snippets, so that
1070 the full configuration is shown.
1072 * systemctl now supports globbing on the various "list-xyz"
1073 commands, like "list-units" or "list-sockets", as well as on
1074 those commands which take multiple unit names.
1076 * journalctl's --unit= switch gained support for globbing.
1078 * All systemd daemons now make use of the watchdog logic so
1079 that systemd automatically notices when they hang.
1081 * If the $container_ttys environment variable is set,
1082 getty-generator will automatically spawn a getty for each
1083 listed tty. This is useful for container managers to request
1084 login gettys to be spawned on as many ttys as needed.
1086 * %h, %s, %U specifier support is not available anymore when
1087 used in unit files for PID 1. This is because NSS calls are
1088 not safe from PID 1. They stay available for --user
1089 instances of systemd, and as special case for the root user.
1091 * loginctl gained a new "--no-legend" switch to turn off output
1094 * The "sd-login.h" API gained three new calls:
1095 sd_session_is_remote(), sd_session_get_remote_user(),
1096 sd_session_get_remote_host() to query information about
1099 * The udev hardware database now also carries vendor/product
1100 information of SDIO devices.
1102 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
1103 determine whether watchdog notifications are requested by
1106 * Socket-activated per-connection services now include a
1107 short description of the connection parameters in the
1110 * tmpfiles gained a new "--boot" option. When this is not used,
1111 only lines where the command character is not suffixed with
1112 "!" are executed. When this option is specified, those
1113 options are executed too. This partitions tmpfiles
1114 directives into those that can be safely executed at any
1115 time, and those which should be run only at boot (for
1116 example, a line that creates /run/nologin).
1118 * A new API "sd-resolve.h" has been added which provides a simple
1119 asynchronous wrapper around glibc NSS host name resolution
1120 calls, such as getaddrinfo(). In contrast to glibc's
1121 getaddrinfo_a(), it does not use signals. In contrast to most
1122 other asynchronous name resolution libraries, this one does
1123 not reimplement DNS, but reuses NSS, so that alternate
1124 host name resolution systems continue to work, such as mDNS,
1125 LDAP, etc. This API is based on libasyncns, but it has been
1126 cleaned up for inclusion in systemd.
1128 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
1129 "sd-daemon.h" are no longer found in individual libraries
1130 libsystemd-journal.so, libsystemd-login.so,
1131 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
1132 merged them into a single library, libsystemd.so, which
1133 provides all symbols. The reason for this is cyclic
1134 dependencies, as these libraries tend to use each other's
1135 symbols. So far, we have managed to workaround that by linking
1136 a copy of a good part of our code into each of these
1137 libraries again and again, which, however, makes certain
1138 things hard to do, like sharing static variables. Also, it
1139 substantially increases footprint. With this change, there
1140 is only one library for the basic APIs systemd
1141 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
1142 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
1143 library as well, however are subject to the --enable-kdbus
1144 switch (see below). Note that "sd-dhcp-client.h" is not part
1145 of this library (this is because it only consumes, never
1146 provides, services of/to other APIs). To make the transition
1147 easy from the separate libraries to the unified one, we
1148 provide the --enable-compat-libs compile-time switch which
1149 will generate stub libraries that are compatible with the
1150 old ones but redirect all calls to the new one.
1152 * All of the kdbus logic and the new APIs "sd-bus.h",
1153 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
1154 and "sd-utf8.h" are compile-time optional via the
1155 "--enable-kdbus" switch, and they are not compiled in by
1156 default. To make use of kdbus, you have to explicitly enable
1157 the switch. Note however, that neither the kernel nor the
1158 userspace API for all of this is considered stable yet. We
1159 want to maintain the freedom to still change the APIs for
1160 now. By specifying this build-time switch, you acknowledge
1161 that you are aware of the instability of the current
1164 * Also, note that while kdbus is pretty much complete,
1165 it lacks one thing: proper policy support. This means you
1166 can build a fully working system with all features; however,
1167 it will be highly insecure. Policy support will be added in
1168 one of the next releases, at the same time that we will
1169 declare the APIs stable.
1171 * When the kernel command-line argument "kdbus" is specified,
1172 systemd will automatically load the kdbus.ko kernel module. At
1173 this stage of development, it is only useful for testing kdbus
1174 and should not be used in production. Note: if "--enable-kdbus"
1175 is specified, and the kdbus.ko kernel module is available, and
1176 "kdbus" is added to the kernel command line, the entire system
1177 runs with kdbus instead of dbus-daemon, with the above mentioned
1178 problem of missing the system policy enforcement. Also a future
1179 version of kdbus.ko or a newer systemd will not be compatible with
1180 each other, and will unlikely be able to boot the machine if only
1181 one of them is updated.
1183 * systemctl gained a new "import-environment" command which
1184 uploads the caller's environment (or parts thereof) into the
1185 service manager so that it is inherited by services started
1186 by the manager. This is useful to upload variables like
1187 $DISPLAY into the user service manager.
1189 * A new PrivateDevices= switch has been added to service units
1190 which allows running a service with a namespaced /dev
1191 directory that does not contain any device nodes for
1192 physical devices. More specifically, it only includes devices
1193 such as /dev/null, /dev/urandom, and /dev/zero which are API
1196 * logind has been extended to support behaviour like VT
1197 switching on seats that do not support a VT. This makes
1198 multi-session available on seats that are not the first seat
1199 (seat0), and on systems where kernel support for VTs has
1200 been disabled at compile-time.
1202 * If a process holds a delay lock for system sleep or shutdown
1203 and fails to release it in time, we will now log its
1204 identity. This makes it easier to identify processes that
1205 cause slow suspends or power-offs.
1207 * When parsing /etc/crypttab, support for a new key-slot=
1208 option as supported by Debian is added. It allows indicating
1209 which LUKS slot to use on disk, speeding up key loading.
1211 * The sd_journald_sendv() API call has been checked and
1212 officially declared to be async-signal-safe so that it may
1213 be invoked from signal handlers for logging purposes.
1215 * Boot-time status output is now enabled automatically after a
1216 short timeout if boot does not progress, in order to give
1217 the user an indication what she or he is waiting for.
1219 * The boot-time output has been improved to show how much time
1220 remains until jobs expire.
1222 * The KillMode= switch in service units gained a new possible
1223 value "mixed". If set, and the unit is shut down, then the
1224 initial SIGTERM signal is sent only to the main daemon
1225 process, while the following SIGKILL signal is sent to
1226 all remaining processes of the service.
1228 * When a scope unit is registered, a new property "Controller"
1229 may be set. If set to a valid bus name, systemd will send a
1230 RequestStop() signal to this name when it would like to shut
1231 down the scope. This may be used to hook manager logic into
1232 the shutdown logic of scope units. Also, scope units may now
1233 be put in a special "abandoned" state, in which case the
1234 manager process which created them takes no further
1235 responsibilities for it.
1237 * When reading unit files, systemd will now verify
1238 the access mode of these files, and warn about certain
1239 suspicious combinations. This has been added to make it
1240 easier to track down packaging bugs where unit files are
1241 marked executable or world-writable.
1243 * systemd-nspawn gained a new "--setenv=" switch to set
1244 container-wide environment variables. The similar option in
1245 systemd-activate was renamed from "--environment=" to
1246 "--setenv=" for consistency.
1248 * systemd-nspawn has been updated to create a new kdbus domain
1249 for each container that is invoked, thus allowing each
1250 container to have its own set of system and user buses,
1251 independent of the host.
1253 * systemd-nspawn gained a new --drop-capability= switch to run
1254 the container with less capabilities than the default. Both
1255 --drop-capability= and --capability= now take the special
1256 string "all" for dropping or keeping all capabilities.
1258 * systemd-nspawn gained new switches for executing containers
1259 with specific SELinux labels set.
1261 * systemd-nspawn gained a new --quiet switch to not generate
1262 any additional output but the container's own console
1265 * systemd-nspawn gained a new --share-system switch to run a
1266 container without PID namespacing enabled.
1268 * systemd-nspawn gained a new --register= switch to control
1269 whether the container is registered with systemd-machined or
1270 not. This is useful for containers that do not run full
1271 OS images, but only specific apps.
1273 * systemd-nspawn gained a new --keep-unit which may be used
1274 when invoked as the only program from a service unit, and
1275 results in registration of the unit service itself in
1276 systemd-machined, instead of a newly opened scope unit.
1278 * systemd-nspawn gained a new --network-interface= switch for
1279 moving arbitrary interfaces to the container. The new
1280 --network-veth switch creates a virtual Ethernet connection
1281 between host and container. The new --network-bridge=
1282 switch then allows assigning the host side of this virtual
1283 Ethernet connection to a bridge device.
1285 * systemd-nspawn gained a new --personality= switch for
1286 setting the kernel personality for the container. This is
1287 useful when running a 32-bit container on a 64-bit host. A
1288 similar option Personality= is now also available for service
1291 * logind will now also track a "Desktop" identifier for each
1292 session which encodes the desktop environment of it. This is
1293 useful for desktop environments that want to identify
1294 multiple running sessions of itself easily.
1296 * A new SELinuxContext= setting for service units has been
1297 added that allows setting a specific SELinux execution
1298 context for a service.
1300 * Most systemd client tools will now honour $SYSTEMD_LESS for
1301 settings of the "less" pager. By default, these tools will
1302 override $LESS to allow certain operations to work, such as
1303 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
1304 influence this logic.
1306 * systemd's "seccomp" hook-up has been changed to make use of
1307 the libseccomp library instead of using its own
1308 implementation. This has benefits for portability among
1311 * For usage together with SystemCallFilter=, a new
1312 SystemCallErrorNumber= setting has been introduced that
1313 allows configuration of a system error number to be returned
1314 on filtered system calls, instead of immediately killing the
1315 process. Also, SystemCallArchitectures= has been added to
1316 limit access to system calls of a particular architecture
1317 (in order to turn off support for unused secondary
1318 architectures). There is also a global
1319 SystemCallArchitectures= setting in system.conf now to turn
1320 off support for non-native system calls system-wide.
1322 * systemd requires a kernel with a working name_to_handle_at(),
1323 please see the kernel config requirements in the README file.
1325 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
1326 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
1327 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
1328 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
1329 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
1330 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
1331 Elia Pinto, Florian Weimer, George McCollister, Goffredo
1332 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
1333 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
1334 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
1335 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
1336 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
1337 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
1338 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
1339 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
1340 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
1341 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
1342 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
1343 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
1344 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
1345 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
1346 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
1347 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
1348 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
1350 -- Berlin, 2014-02-20
1354 * logind has gained support for facilitating privileged input
1355 and drm device access for unprivileged clients. This work is
1356 useful to allow Wayland display servers (and similar
1357 programs, such as kmscon) to run under the user's ID and
1358 access input and drm devices which are normally
1359 protected. When this is used (and the kernel is new enough)
1360 logind will "mute" IO on the file descriptors passed to
1361 Wayland as long as it is in the background and "unmute" it
1362 if it returns into the foreground. This allows secure
1363 session switching without allowing background sessions to
1364 eavesdrop on input and display data. This also introduces
1365 session switching support if VT support is turned off in the
1366 kernel, and on seats that are not seat0.
1368 * A new kernel command line option luks.options= is understood
1369 now which allows specifiying LUKS options for usage for LUKS
1370 encrypted partitions specified with luks.uuid=.
1372 * tmpfiles.d(5) snippets may now use specifier expansion in
1373 path names. More specifically %m, %b, %H, %v, are now
1374 replaced by the local machine id, boot id, hostname, and
1375 kernel version number.
1377 * A new tmpfiles.d(5) command "m" has been introduced which
1378 may be used to change the owner/group/access mode of a file
1379 or directory if it exists, but do nothing if it does not.
1381 * This release removes high-level support for the
1382 MemorySoftLimit= cgroup setting. The underlying kernel
1383 cgroup attribute memory.soft_limit= is currently badly
1384 designed and likely to be removed from the kernel API in its
1385 current form, hence we should not expose it for now.
1387 * The memory.use_hierarchy cgroup attribute is now enabled for
1388 all cgroups systemd creates in the memory cgroup
1389 hierarchy. This option is likely to be come the built-in
1390 default in the kernel anyway, and the non-hierarchial mode
1391 never made much sense in the intrinsically hierarchial
1394 * A new field _SYSTEMD_SLICE= is logged along with all journal
1395 messages containing the slice a message was generated
1396 from. This is useful to allow easy per-customer filtering of
1397 logs among other things.
1399 * systemd-journald will no longer adjust the group of journal
1400 files it creates to the "systemd-journal" group. Instead we
1401 rely on the journal directory to be owned by the
1402 "systemd-journal" group, and its setgid bit set, so that the
1403 kernel file system layer will automatically enforce that
1404 journal files inherit this group assignment. The reason for
1405 this change is that we cannot allow NSS look-ups from
1406 journald which would be necessary to resolve
1407 "systemd-journal" to a numeric GID, because this might
1408 create deadlocks if NSS involves synchronous queries to
1409 other daemons (such as nscd, or sssd) which in turn are
1410 logging clients of journald and might block on it, which
1411 would then dead lock. A tmpfiles.d(5) snippet included in
1412 systemd will make sure the setgid bit and group are
1413 properly set on the journal directory if it exists on every
1414 boot. However, we recommend adjusting it manually after
1415 upgrades too (or from RPM scriptlets), so that the change is
1416 not delayed until next reboot.
1418 * Backlight and random seed files in /var/lib/ have moved into
1419 the /var/lib/systemd/ directory, in order to centralize all
1420 systemd generated files in one directory.
1422 * Boot time performance measurements (as displayed by
1423 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
1424 performance information if that's available to determine how
1425 much time BIOS and boot loader initialization required. With
1426 a sufficiently new BIOS you hence no longer need to boot
1427 with Gummiboot to get access to such information.
1429 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
1430 Cristian Rodríguez, Dave Reisner, David Herrmann, David
1431 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
1432 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
1433 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
1434 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
1435 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1437 -- Berlin, 2013-10-02
1441 * The Restart= option for services now understands a new
1442 on-watchdog setting, which will restart the service
1443 automatically if the service stops sending out watchdog keep
1444 alive messages (as configured with WatchdogSec=).
1446 * The getty generator (which is responsible for bringing up a
1447 getty on configured serial consoles) will no longer only
1448 start a getty on the primary kernel console but on all
1449 others, too. This makes the order in which console= is
1450 specified on the kernel command line less important.
1452 * libsystemd-logind gained a new sd_session_get_vt() call to
1453 retrieve the VT number of a session.
1455 * If the option "tries=0" is set for an entry of /etc/crypttab
1456 its passphrase is queried indefinitely instead of any
1457 maximum number of tries.
1459 * If a service with a configure PID file terminates its PID
1460 file will now be removed automatically if it still exists
1461 afterwards. This should put an end to stale PID files.
1463 * systemd-run will now also take relative binary path names
1464 for execution and no longer insists on absolute paths.
1466 * InaccessibleDirectories= and ReadOnlyDirectories= now take
1467 paths that are optionally prefixed with "-" to indicate that
1468 it should not be considered a failure if they do not exist.
1470 * journalctl -o (and similar commands) now understands a new
1471 output mode "short-precise", it is similar to "short" but
1472 shows timestamps with usec accuracy.
1474 * The option "discard" (as known from Debian) is now
1475 synonymous to "allow-discards" in /etc/crypttab. In fact,
1476 "discard" is preferred now (since it is easier to remember
1479 * Some licensing clean-ups were made, so that more code is now
1480 LGPL-2.1 licensed than before.
1482 * A minimal tool to save/restore the display backlight
1483 brightness across reboots has been added. It will store the
1484 backlight setting as late as possible at shutdown, and
1485 restore it as early as possible during reboot.
1487 * A logic to automatically discover and enable home and swap
1488 partitions on GPT disks has been added. With this in place
1489 /etc/fstab becomes optional for many setups as systemd can
1490 discover certain partitions located on the root disk
1491 automatically. Home partitions are recognized under their
1492 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
1493 partitions are recognized under their GPT type ID
1494 0657fd6da4ab43c484e50933c84b4f4f.
1496 * systemd will no longer pass any environment from the kernel
1497 or initrd to system services. If you want to set an
1498 environment for all services, do so via the kernel command
1499 line systemd.setenv= assignment.
1501 * The systemd-sysctl tool no longer natively reads the file
1502 /etc/sysctl.conf. If desired, the file should be symlinked
1503 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
1504 legacy support by a symlink rather than built-in code, it
1505 also makes the otherwise hidden order of application of the
1506 different files visible. (Note that this partly reverts to a
1507 pre-198 application order of sysctl knobs!)
1509 * The "systemctl set-log-level" and "systemctl dump" commands
1510 have been moved to systemd-analyze.
1512 * systemd-run learned the new --remain-after-exit switch,
1513 which causes the scope unit not to be cleaned up
1514 automatically after the process terminated.
1516 * tmpfiles learned a new --exclude-prefix= switch to exclude
1517 certain paths from operation.
1519 * journald will now automatically flush all messages to disk
1520 as soon as a message of the log priorities CRIT, ALERT or
1523 Contributions from: Andrew Cook, Brandon Philips, Christian
1524 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
1525 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
1526 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
1527 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
1528 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
1529 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
1530 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
1531 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
1532 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
1533 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
1534 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
1535 William Giokas, Zbigniew Jędrzejewski-Szmek
1537 -- Berlin, 2013-09-13
1541 * The documentation has been updated to cover the various new
1542 concepts introduced with 205.
1544 * Unit files now understand the new %v specifier which
1545 resolves to the kernel version string as returned by "uname
1548 * systemctl now supports filtering the unit list output by
1549 load state, active state and sub state, using the new
1552 * "systemctl status" will now show the results of the
1553 condition checks (like ConditionPathExists= and similar) of
1554 the last start attempts of the unit. They are also logged to
1557 * "journalctl -b" may now be used to look for boot output of a
1558 specific boot. Try "journalctl -b -1" for the previous boot,
1559 but the syntax is substantially more powerful.
1561 * "journalctl --show-cursor" has been added which prints the
1562 cursor string the last shown log line. This may then be used
1563 with the new "journalctl --after-cursor=" switch to continue
1564 browsing logs from that point on.
1566 * "journalctl --force" may now be used to force regeneration
1569 * Creation of "dead" device nodes has been moved from udev
1570 into kmod and tmpfiles. Previously, udev would read the kmod
1571 databases to pre-generate dead device nodes based on meta
1572 information contained in kernel modules, so that these would
1573 be auto-loaded on access rather then at boot. As this
1574 does not really have much to do with the exposing actual
1575 kernel devices to userspace this has always been slightly
1576 alien in the udev codebase. Following the new scheme kmod
1577 will now generate a runtime snippet for tmpfiles from the
1578 module meta information and it now is tmpfiles' job to the
1579 create the nodes. This also allows overriding access and
1580 other parameters for the nodes using the usual tmpfiles
1581 facilities. As side effect this allows us to remove the
1582 CAP_SYS_MKNOD capability bit from udevd entirely.
1584 * logind's device ACLs may now be applied to these "dead"
1585 devices nodes too, thus finally allowing managed access to
1586 devices such as /dev/snd/sequencer whithout loading the
1587 backing module right-away.
1589 * A new RPM macro has been added that may be used to apply
1590 tmpfiles configuration during package installation.
1592 * systemd-detect-virt and ConditionVirtualization= now can
1593 detect User-Mode-Linux machines (UML).
1595 * journald will now implicitly log the effective capabilities
1596 set of processes in the message metadata.
1598 * systemd-cryptsetup has gained support for TrueCrypt volumes.
1600 * The initrd interface has been simplified (more specifically,
1601 support for passing performance data via environment
1602 variables and fsck results via files in /run has been
1603 removed). These features were non-essential, and are
1604 nowadays available in a much nicer way by having systemd in
1605 the initrd serialize its state and have the hosts systemd
1606 deserialize it again.
1608 * The udev "keymap" data files and tools to apply keyboard
1609 specific mappings of scan to key codes, and force-release
1610 scan code lists have been entirely replaced by a udev
1611 "keyboard" builtin and a hwdb data file.
1613 * systemd will now honour the kernel's "quiet" command line
1614 argument also during late shutdown, resulting in a
1615 completely silent shutdown when used.
1617 * There's now an option to control the SO_REUSEPORT socket
1618 option in .socket units.
1620 * Instance units will now automatically get a per-template
1621 subslice of system.slice unless something else is explicitly
1622 configured. For example, instances of sshd@.service will now
1623 implicitly be placed in system-sshd.slice rather than
1624 system.slice as before.
1626 * Test coverage support may now be enabled at build time.
1628 Contributions from: Dave Reisner, Frederic Crozat, Harald
1629 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
1630 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
1631 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
1632 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
1633 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
1634 Giokas, Zbigniew Jędrzejewski-Szmek
1636 -- Berlin, 2013-07-23
1640 * Two new unit types have been introduced:
1642 Scope units are very similar to service units, however, are
1643 created out of pre-existing processes -- instead of PID 1
1644 forking off the processes. By using scope units it is
1645 possible for system services and applications to group their
1646 own child processes (worker processes) in a powerful way
1647 which then maybe used to organize them, or kill them
1648 together, or apply resource limits on them.
1650 Slice units may be used to partition system resources in an
1651 hierarchial fashion and then assign other units to them. By
1652 default there are now three slices: system.slice (for all
1653 system services), user.slice (for all user sessions),
1654 machine.slice (for VMs and containers).
1656 Slices and scopes have been introduced primarily in
1657 context of the work to move cgroup handling to a
1658 single-writer scheme, where only PID 1
1659 creates/removes/manages cgroups.
1661 * There's a new concept of "transient" units. In contrast to
1662 normal units these units are created via an API at runtime,
1663 not from configuration from disk. More specifically this
1664 means it is now possible to run arbitrary programs as
1665 independent services, with all execution parameters passed
1666 in via bus APIs rather than read from disk. Transient units
1667 make systemd substantially more dynamic then it ever was,
1668 and useful as a general batch manager.
1670 * logind has been updated to make use of scope and slice units
1671 for managing user sessions. As a user logs in he will get
1672 his own private slice unit, to which all sessions are added
1673 as scope units. We also added support for automatically
1674 adding an instance of user@.service for the user into the
1675 slice. Effectively logind will no longer create cgroup
1676 hierarchies on its own now, it will defer entirely to PID 1
1677 for this by means of scope, service and slice units. Since
1678 user sessions this way become entities managed by PID 1
1679 the output of "systemctl" is now a lot more comprehensive.
1681 * A new mini-daemon "systemd-machined" has been added which
1682 may be used by virtualization managers to register local
1683 VMs/containers. nspawn has been updated accordingly, and
1684 libvirt will be updated shortly. machined will collect a bit
1685 of meta information about the VMs/containers, and assign
1686 them their own scope unit (see above). The collected
1687 meta-data is then made available via the "machinectl" tool,
1688 and exposed in "ps" and similar tools. machined/machinectl
1689 is compile-time optional.
1691 * As discussed earlier, the low-level cgroup configuration
1692 options ControlGroup=, ControlGroupModify=,
1693 ControlGroupPersistent=, ControlGroupAttribute= have been
1694 removed. Please use high-level attribute settings instead as
1695 well as slice units.
1697 * A new bus call SetUnitProperties() has been added to alter
1698 various runtime parameters of a unit. This is primarily
1699 useful to alter cgroup parameters dynamically in a nice way,
1700 but will be extended later on to make more properties
1701 modifiable at runtime. systemctl gained a new set-properties
1702 command that wraps this call.
1704 * A new tool "systemd-run" has been added which can be used to
1705 run arbitrary command lines as transient services or scopes,
1706 while configuring a number of settings via the command
1707 line. This tool is currently very basic, however already
1708 very useful. We plan to extend this tool to even allow
1709 queuing of execution jobs with time triggers from the
1710 command line, similar in fashion to "at".
1712 * nspawn will now inform the user explicitly that kernels with
1713 audit enabled break containers, and suggest the user to turn
1716 * Support for detecting the IMA and AppArmor security
1717 frameworks with ConditionSecurity= has been added.
1719 * journalctl gained a new "-k" switch for showing only kernel
1720 messages, mimicking dmesg output; in addition to "--user"
1721 and "--system" switches for showing only user's own logs
1724 * systemd-delta can now show information about drop-in
1725 snippets extending unit files.
1727 * libsystemd-bus has been substantially updated but is still
1728 not available as public API.
1730 * systemd will now look for the "debug" argument on the kernel
1731 command line and enable debug logging, similar to what
1732 "systemd.log_level=debug" already did before.
1734 * "systemctl set-default", "systemctl get-default" has been
1735 added to configure the default.target symlink, which
1736 controls what to boot into by default.
1738 * "systemctl set-log-level" has been added as a convenient
1739 way to raise and lower systemd logging threshold.
1741 * "systemd-analyze plot" will now show the time the various
1742 generators needed for execution, as well as information
1743 about the unit file loading.
1745 * libsystemd-journal gained a new sd_journal_open_files() call
1746 for opening specific journal files. journactl also gained a
1747 new switch to expose this new functionality. Previously we
1748 only supported opening all files from a directory, or all
1749 files from the system, as opening individual files only is
1750 racy due to journal file rotation.
1752 * systemd gained the new DefaultEnvironment= setting in
1753 /etc/systemd/system.conf to set environment variables for
1756 * If a privileged process logs a journal message with the
1757 OBJECT_PID= field set, then journald will automatically
1758 augment this with additional OBJECT_UID=, OBJECT_GID=,
1759 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
1760 system services want to log events about specific client
1761 processes. journactl/systemctl has been updated to make use
1762 of this information if all log messages regarding a specific
1765 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
1766 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
1767 Reisner, David Coppa, David King, David Strauss, Eelco
1768 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
1769 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
1770 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
1771 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
1772 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
1773 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
1774 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
1775 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
1776 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
1777 Łukasz Stelmach, 장동준
1781 * The Python bindings gained some minimal support for the APIs
1782 exposed by libsystemd-logind.
1784 * ConditionSecurity= gained support for detecting SMACK. Since
1785 this condition already supports SELinux and AppArmor we only
1786 miss IMA for this. Patches welcome!
1788 Contributions from: Karol Lewandowski, Lennart Poettering,
1789 Zbigniew Jędrzejewski-Szmek
1793 * systemd-nspawn will now create /etc/resolv.conf if
1794 necessary, before bind-mounting the host's file onto it.
1796 * systemd-nspawn will now store meta information about a
1797 container on the container's cgroup as extended attribute
1798 fields, including the root directory.
1800 * The cgroup hierarchy has been reworked in many ways. All
1801 objects any of the components systemd creates in the cgroup
1802 tree are now suffixed. More specifically, user sessions are
1803 now placed in cgroups suffixed with ".session", users in
1804 cgroups suffixed with ".user", and nspawn containers in
1805 cgroups suffixed with ".nspawn". Furthermore, all cgroup
1806 names are now escaped in a simple scheme to avoid collision
1807 of userspace object names with kernel filenames. This work
1808 is preparation for making these objects relocatable in the
1809 cgroup tree, in order to allow easy resource partitioning of
1810 these objects without causing naming conflicts.
1812 * systemctl list-dependencies gained the new switches
1813 --plain, --reverse, --after and --before.
1815 * systemd-inhibit now shows the process name of processes that
1816 have taken an inhibitor lock.
1818 * nss-myhostname will now also resolve "localhost"
1819 implicitly. This makes /etc/hosts an optional file and
1820 nicely handles that on IPv6 ::1 maps to both "localhost" and
1823 * libsystemd-logind.so gained a new call
1824 sd_get_machine_names() to enumerate running containers and
1825 VMs (currently only supported by very new libvirt and
1826 nspawn). sd_login_monitor can now be used to watch
1827 VMs/containers coming and going.
1829 * .include is not allowed recursively anymore, and only in
1830 unit files. Usually it is better to use drop-in snippets in
1831 .d/*.conf anyway, as introduced with systemd 198.
1833 * systemd-analyze gained a new "critical-chain" command that
1834 determines the slowest chain of units run during system
1835 boot-up. It is very useful for tracking down where
1836 optimizing boot time is the most beneficial.
1838 * systemd will no longer allow manipulating service paths in
1839 the name=systemd:/system cgroup tree using ControlGroup= in
1840 units. (But is still fine with it in all other dirs.)
1842 * There's a new systemd-nspawn@.service service file that may
1843 be used to easily run nspawn containers as system
1844 services. With the container's root directory in
1845 /var/lib/container/foobar it is now sufficient to run
1846 "systemctl start systemd-nspawn@foobar.service" to boot it.
1848 * systemd-cgls gained a new parameter "--machine" to list only
1849 the processes within a certain container.
1851 * ConditionSecurity= now can check for "apparmor". We still
1852 are lacking checks for SMACK and IMA for this condition
1853 check though. Patches welcome!
1855 * A new configuration file /etc/systemd/sleep.conf has been
1856 added that may be used to configure which kernel operation
1857 systemd is supposed to execute when "suspend", "hibernate"
1858 or "hybrid-sleep" is requested. This makes the new kernel
1859 "freeze" state accessible to the user.
1861 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
1862 the passed argument if applicable.
1864 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
1865 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
1866 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
1867 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
1868 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
1869 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
1870 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
1875 * The output of 'systemctl list-jobs' got some polishing. The
1876 '--type=' argument may now be passed more than once. A new
1877 command 'systemctl list-sockets' has been added which shows
1878 a list of kernel sockets systemd is listening on with the
1879 socket units they belong to, plus the units these socket
1882 * The experimental libsystemd-bus library got substantial
1883 updates to work in conjunction with the (also experimental)
1884 kdbus kernel project. It works well enough to exchange
1885 messages with some sophistication. Note that kdbus is not
1886 ready yet, and the library is mostly an elaborate test case
1887 for now, and not installable.
1889 * systemd gained a new unit 'systemd-static-nodes.service'
1890 that generates static device nodes earlier during boot, and
1891 can run in conjunction with udev.
1893 * libsystemd-login gained a new call sd_pid_get_user_unit()
1894 to retrieve the user systemd unit a process is running
1895 in. This is useful for systems where systemd is used as
1898 * systemd-nspawn now places all containers in the new /machine
1899 top-level cgroup directory in the name=systemd
1900 hierarchy. libvirt will soon do the same, so that we get a
1901 uniform separation of /system, /user and /machine for system
1902 services, user processes and containers/virtual
1903 machines. This new cgroup hierarchy is also useful to stick
1904 stable names to specific container instances, which can be
1905 recognized later this way (this name may be controlled
1906 via systemd-nspawn's new -M switch). libsystemd-login also
1907 gained a new call sd_pid_get_machine_name() to retrieve the
1908 name of the container/VM a specific process belongs to.
1910 * bootchart can now store its data in the journal.
1912 * libsystemd-journal gained a new call
1913 sd_journal_add_conjunction() for AND expressions to the
1914 matching logic. This can be used to express more complex
1915 logical expressions.
1917 * journactl can now take multiple --unit= and --user-unit=
1920 * The cryptsetup logic now understands the "luks.key=" kernel
1921 command line switch for specifying a file to read the
1922 decryption key from. Also, if a configured key file is not
1923 found the tool will now automatically fall back to prompting
1926 * Python systemd.journal module was updated to wrap recently
1927 added functions from libsystemd-journal. The interface was
1928 changed to bring the low level interface in s.j._Reader
1929 closer to the C API, and the high level interface in
1930 s.j.Reader was updated to wrap and convert all data about
1933 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
1934 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
1935 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
1936 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
1937 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
1938 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1942 * journalctl --update-catalog now understands a new --root=
1943 option to operate on catalogs found in a different root
1946 * During shutdown after systemd has terminated all running
1947 services a final killing loop kills all remaining left-over
1948 processes. We will now print the name of these processes
1949 when we send SIGKILL to them, since this usually indicates a
1952 * If /etc/crypttab refers to password files stored on
1953 configured mount points automatic dependencies will now be
1954 generated to ensure the specific mount is established first
1955 before the key file is attempted to be read.
1957 * 'systemctl status' will now show information about the
1958 network sockets a socket unit is listening on.
1960 * 'systemctl status' will also shown information about any
1961 drop-in configuration file for units. (Drop-In configuration
1962 files in this context are files such as
1963 /etc/systemd/systemd/foobar.service.d/*.conf)
1965 * systemd-cgtop now optionally shows summed up CPU times of
1966 cgroups. Press '%' while running cgtop to switch between
1967 percentage and absolute mode. This is useful to determine
1968 which cgroups use up the most CPU time over the entire
1969 runtime of the system. systemd-cgtop has also been updated
1970 to be 'pipeable' for processing with further shell tools.
1972 * 'hostnamectl set-hostname' will now allow setting of FQDN
1975 * The formatting and parsing of time span values has been
1976 changed. The parser now understands fractional expressions
1977 such as "5.5h". The formatter will now output fractional
1978 expressions for all time spans under 1min, i.e. "5.123456s"
1979 rather than "5s 123ms 456us". For time spans under 1s
1980 millisecond values are shown, for those under 1ms
1981 microsecond values are shown. This should greatly improve
1982 all time-related output of systemd.
1984 * libsystemd-login and libsystemd-journal gained new
1985 functions for querying the poll() events mask and poll()
1986 timeout value for integration into arbitrary event
1989 * localectl gained the ability to list available X11 keymaps
1990 (models, layouts, variants, options).
1992 * 'systemd-analyze dot' gained the ability to filter for
1993 specific units via shell-style globs, to create smaller,
1994 more useful graphs. I.e. it is now possible to create simple
1995 graphs of all the dependencies between only target units, or
1996 of all units that Avahi has dependencies with.
1998 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
1999 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
2000 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
2001 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
2002 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
2003 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
2004 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
2008 * The boot-time readahead implementation for rotating media
2009 will now read the read-ahead data in multiple passes which
2010 consist of all read requests made in equidistant time
2011 intervals. This means instead of strictly reading read-ahead
2012 data in its physical order on disk we now try to find a
2013 middle ground between physical and access time order.
2015 * /etc/os-release files gained a new BUILD_ID= field for usage
2016 on operating systems that provide continuous builds of OS
2019 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
2020 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
2021 William Douglas, Zbigniew Jędrzejewski-Szmek
2025 * systemd-python gained an API exposing libsystemd-daemon.
2027 * The SMACK setup logic gained support for uploading CIPSO
2030 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
2031 ReadOnlyDirectories= and InaccessibleDirectories= has
2032 changed. The private /tmp and /var/tmp directories are now
2033 shared by all processes of a service (which means
2034 ExecStartPre= may now leave data in /tmp that ExecStart= of
2035 the same service can still access). When a service is
2036 stopped its temporary directories are immediately deleted
2037 (normal clean-up with tmpfiles is still done in addition to
2040 * By default, systemd will now set a couple of sysctl
2041 variables in the kernel: the safe sysrq options are turned
2042 on, IP route verification is turned on, and source routing
2043 disabled. The recently added hardlink and softlink
2044 protection of the kernel is turned on. These settings should
2045 be reasonably safe, and good defaults for all new systems.
2047 * The predictable network naming logic may now be turned off
2048 with a new kernel command line switch: net.ifnames=0.
2050 * A new libsystemd-bus module has been added that implements a
2051 pretty complete D-Bus client library. For details see:
2053 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
2055 * journald will now explicitly flush the journal files to disk
2056 at the latest 5min after each write. The file will then also
2057 be marked offline until the next write. This should increase
2058 reliability in case of a crash. The synchronization delay
2059 can be configured via SyncIntervalSec= in journald.conf.
2061 * There's a new remote-fs-setup.target unit that can be used
2062 to pull in specific services when at least one remote file
2063 system is to be mounted.
2065 * There are new targets timers.target and paths.target as
2066 canonical targets to pull user timer and path units in
2067 from. This complements sockets.target with a similar
2068 purpose for socket units.
2070 * libudev gained a new call udev_device_set_attribute_value()
2071 to set sysfs attributes of a device.
2073 * The udev daemon now sets the default number of worker
2074 processes executed in parallel based on the number of available
2075 CPUs instead of the amount of available RAM. This is supposed
2076 to provide a more reliable default and limit a too aggressive
2077 paralellism for setups with 1000s of devices connected.
2079 Contributions from: Auke Kok, Colin Walters, Cristian
2080 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
2081 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
2082 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
2083 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
2084 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
2085 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
2086 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
2087 Zbigniew Jędrzejewski-Szmek
2091 * Configuration of unit files may now be extended via drop-in
2092 files without having to edit/override the unit files
2093 themselves. More specifically, if the administrator wants to
2094 change one value for a service file foobar.service he can
2095 now do so by dropping in a configuration snippet into
2096 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
2097 will load all these snippets and apply them on top of the
2098 main unit configuration file, possibly extending or
2099 overriding its settings. Using these drop-in snippets is
2100 generally nicer than the two earlier options for changing
2101 unit files locally: copying the files from
2102 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
2103 them there; or creating a new file in /etc/systemd/system/
2104 that incorporates the original one via ".include". Drop-in
2105 snippets into these .d/ directories can be placed in any
2106 directory systemd looks for units in, and the usual
2107 overriding semantics between /usr/lib, /etc and /run apply
2110 * Most unit file settings which take lists of items can now be
2111 reset by assigning the empty string to them. For example,
2112 normally, settings such as Environment=FOO=BAR append a new
2113 environment variable assignment to the environment block,
2114 each time they are used. By assigning Environment= the empty
2115 string the environment block can be reset to empty. This is
2116 particularly useful with the .d/*.conf drop-in snippets
2117 mentioned above, since this adds the ability to reset list
2118 settings from vendor unit files via these drop-ins.
2120 * systemctl gained a new "list-dependencies" command for
2121 listing the dependencies of a unit recursively.
2123 * Inhibitors are now honored and listed by "systemctl
2124 suspend", "systemctl poweroff" (and similar) too, not only
2125 GNOME. These commands will also list active sessions by
2128 * Resource limits (as exposed by the various control group
2129 controllers) can now be controlled dynamically at runtime
2130 for all units. More specifically, you can now use a command
2131 like "systemctl set-cgroup-attr foobar.service cpu.shares
2132 2000" to alter the CPU shares a specific service gets. These
2133 settings are stored persistently on disk, and thus allow the
2134 administrator to easily adjust the resource usage of
2135 services with a few simple commands. This dynamic resource
2136 management logic is also available to other programs via the
2137 bus. Almost any kernel cgroup attribute and controller is
2140 * systemd-vconsole-setup will now copy all font settings to
2141 all allocated VTs, where it previously applied them only to
2144 * libsystemd-login gained the new sd_session_get_tty() API
2147 * This release drops support for a few legacy or
2148 distribution-specific LSB facility names when parsing init
2149 scripts: $x-display-manager, $mail-transfer-agent,
2150 $mail-transport-agent, $mail-transfer-agent, $smtp,
2151 $null. Also, the mail-transfer-agent.target unit backing
2152 this has been removed. Distributions which want to retain
2153 compatibility with this should carry the burden for
2154 supporting this themselves and patch support for these back
2155 in, if they really need to. Also, the facilities $syslog and
2156 $local_fs are now ignored, since systemd does not support
2157 early-boot LSB init scripts anymore, and these facilities
2158 are implied anyway for normal services. syslog.target has
2161 * There are new bus calls on PID1's Manager object for
2162 cancelling jobs, and removing snapshot units. Previously,
2163 both calls were only available on the Job and Snapshot
2166 * systemd-journal-gatewayd gained SSL support.
2168 * The various "environment" files, such as /etc/locale.conf
2169 now support continuation lines with a backslash ("\") as
2170 last character in the line, similarly in style (but different)
2171 to how this is supported in shells.
2173 * For normal user processes the _SYSTEMD_USER_UNIT= field is
2174 now implicitly appended to every log entry logged. systemctl
2175 has been updated to filter by this field when operating on a
2176 user systemd instance.
2178 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
2179 CAP_AUDIT_CONTROL capabilities to the capabilities set for
2180 the container. This makes it easier to boot unmodified
2181 Fedora systems in a container, which however still requires
2182 audit=0 to be passed on the kernel command line. Auditing in
2183 kernel and userspace is unfortunately still too broken in
2184 context of containers, hence we recommend compiling it out
2185 of the kernel or using audit=0. Hopefully this will be fixed
2186 one day for good in the kernel.
2188 * nspawn gained the new --bind= and --bind-ro= parameters to
2189 bind mount specific directories from the host into the
2192 * nspawn will now mount its own devpts file system instance
2193 into the container, in order not to leak pty devices from
2194 the host into the container.
2196 * systemd will now read the firmware boot time performance
2197 information from the EFI variables, if the used boot loader
2198 supports this, and takes it into account for boot performance
2199 analysis via "systemd-analyze". This is currently supported
2200 only in conjunction with Gummiboot, but could be supported
2201 by other boot loaders too. For details see:
2203 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
2205 * A new generator has been added that automatically mounts the
2206 EFI System Partition (ESP) to /boot, if that directory
2207 exists, is empty, and no other file system has been
2208 configured to be mounted there.
2210 * logind will now send out PrepareForSleep(false) out
2211 unconditionally, after coming back from suspend. This may be
2212 used by applications as asynchronous notification for
2213 system resume events.
2215 * "systemctl unlock-sessions" has been added, that allows
2216 unlocking the screens of all user sessions at once, similar
2217 to how "systemctl lock-sessions" already locked all users
2218 sessions. This is backed by a new D-Bus call UnlockSessions().
2220 * "loginctl seat-status" will now show the master device of a
2221 seat. (i.e. the device of a seat that needs to be around for
2222 the seat to be considered available, usually the graphics
2225 * tmpfiles gained a new "X" line type, that allows
2226 configuration of files and directories (with wildcards) that
2227 shall be excluded from automatic cleanup ("aging").
2229 * udev default rules set the device node permissions now only
2230 at "add" events, and do not change them any longer with a
2231 later "change" event.
2233 * The log messages for lid events and power/sleep keypresses
2234 now carry a message ID.
2236 * We now have a substantially larger unit test suite, but this
2237 continues to be work in progress.
2239 * udevadm hwdb gained a new --root= parameter to change the
2240 root directory to operate relative to.
2242 * logind will now issue a background sync() request to the kernel
2243 early at shutdown, so that dirty buffers are flushed to disk early
2244 instead of at the last moment, in order to optimize shutdown
2247 * A new bootctl tool has been added that is an interface for
2248 certain boot loader operations. This is currently a preview
2249 and is likely to be extended into a small mechanism daemon
2250 like timedated, localed, hostnamed, and can be used by
2251 graphical UIs to enumerate available boot options, and
2252 request boot into firmware operations.
2254 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
2255 the rest of the package. It also has been updated to work
2256 correctly in initrds.
2258 * Policykit previously has been runtime optional, and is now
2259 also compile time optional via a configure switch.
2261 * systemd-analyze has been reimplemented in C. Also "systemctl
2262 dot" has moved into systemd-analyze.
2264 * "systemctl status" with no further parameters will now print
2265 the status of all active or failed units.
2267 * Operations such as "systemctl start" can now be executed
2268 with a new mode "--irreversible" which may be used to queue
2269 operations that cannot accidentally be reversed by a later
2270 job queuing. This is by default used to make shutdown
2271 requests more robust.
2273 * The Python API of systemd now gained a new module for
2274 reading journal files.
2276 * A new tool kernel-install has been added that can install
2277 kernel images according to the Boot Loader Specification:
2279 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
2281 * Boot time console output has been improved to provide
2282 animated boot time output for hanging jobs.
2284 * A new tool systemd-activate has been added which can be used
2285 to test socket activation with, directly from the command
2286 line. This should make it much easier to test and debug
2287 socket activation in daemons.
2289 * journalctl gained a new "--reverse" (or -r) option to show
2290 journal output in reverse order (i.e. newest line first).
2292 * journalctl gained a new "--pager-end" (or -e) option to jump
2293 to immediately jump to the end of the journal in the
2294 pager. This is only supported in conjunction with "less".
2296 * journalctl gained a new "--user-unit=" option, that works
2297 similarly to "--unit=" but filters for user units rather than
2300 * A number of unit files to ease adoption of systemd in
2301 initrds has been added. This moves some minimal logic from
2302 the various initrd implementations into systemd proper.
2304 * The journal files are now owned by a new group
2305 "systemd-journal", which exists specifically to allow access
2306 to the journal, and nothing else. Previously, we used the
2307 "adm" group for that, which however possibly covers more
2308 than just journal/log file access. This new group is now
2309 already used by systemd-journal-gatewayd to ensure this
2310 daemon gets access to the journal files and as little else
2311 as possible. Note that "make install" will also set FS ACLs
2312 up for /var/log/journal to give "adm" and "wheel" read
2313 access to it, in addition to "systemd-journal" which owns
2314 the journal files. We recommend that packaging scripts also
2315 add read access to "adm" + "wheel" to /var/log/journal, and
2316 all existing/future journal files. To normal users and
2317 administrators little changes, however packagers need to
2318 ensure to create the "systemd-journal" system group at
2319 package installation time.
2321 * The systemd-journal-gatewayd now runs as unprivileged user
2322 systemd-journal-gateway:systemd-journal-gateway. Packaging
2323 scripts need to create these system user/group at
2326 * timedated now exposes a new boolean property CanNTP that
2327 indicates whether a local NTP service is available or not.
2329 * systemd-detect-virt will now also detect xen PVs
2331 * The pstore file system is now mounted by default, if it is
2334 * In addition to the SELinux and IMA policies we will now also
2335 load SMACK policies at early boot.
2337 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
2338 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
2339 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
2340 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
2341 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
2342 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
2343 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
2344 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
2345 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
2346 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
2347 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
2348 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
2349 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
2350 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
2354 * Timer units now support calendar time events in addition to
2355 monotonic time events. That means you can now trigger a unit
2356 based on a calendar time specification such as "Thu,Fri
2357 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
2358 or fifth day of any month of the year 2013, given that it is
2359 a thursday or friday. This brings timer event support
2360 considerably closer to cron's capabilities. For details on
2361 the supported calendar time specification language see
2364 * udev now supports a number of different naming policies for
2365 network interfaces for predictable names, and a combination
2366 of these policies is now the default. Please see this wiki
2367 document for details:
2369 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
2371 * Auke Kok's bootchart implementation has been added to the
2372 systemd tree. It is an optional component that can graph the
2373 boot in quite some detail. It is one of the best bootchart
2374 implementations around and minimal in its code and
2377 * nss-myhostname has been integrated into the systemd source
2378 tree. nss-myhostname guarantees that the local hostname
2379 always stays resolvable via NSS. It has been a weak
2380 requirement of systemd-hostnamed since a long time, and
2381 since its code is actually trivial we decided to just
2382 include it in systemd's source tree. It can be turned off
2383 with a configure switch.
2385 * The read-ahead logic is now capable of properly detecting
2386 whether a btrfs file system is on SSD or rotating media, in
2387 order to optimize the read-ahead scheme. Previously, it was
2388 only capable of detecting this on traditional file systems
2391 * In udev, additional device properties are now read from the
2392 IAB in addition to the OUI database. Also, Bluetooth company
2393 identities are attached to the devices as well.
2395 * In service files %U may be used as specifier that is
2396 replaced by the configured user name of the service.
2398 * nspawn may now be invoked without a controlling TTY. This
2399 makes it suitable for invocation as its own service. This
2400 may be used to set up a simple containerized server system
2401 using only core OS tools.
2403 * systemd and nspawn can now accept socket file descriptors
2404 when they are started for socket activation. This enables
2405 implementation of socket activated nspawn
2406 containers. i.e. think about autospawning an entire OS image
2407 when the first SSH or HTTP connection is received. We expect
2408 that similar functionality will also be added to libvirt-lxc
2411 * journalctl will now suppress ANSI color codes when
2412 presenting log data.
2414 * systemctl will no longer show control group information for
2415 a unit if a the control group is empty anyway.
2417 * logind can now automatically suspend/hibernate/shutdown the
2420 * /etc/machine-info and hostnamed now also expose the chassis
2421 type of the system. This can be used to determine whether
2422 the local system is a laptop, desktop, handset or
2423 tablet. This information may either be configured by the
2424 user/vendor or is automatically determined from ACPI and DMI
2425 information if possible.
2427 * A number of PolicyKit actions are now bound together with
2428 "imply" rules. This should simplify creating UIs because
2429 many actions will now authenticate similar ones as well.
2431 * Unit files learnt a new condition ConditionACPower= which
2432 may be used to conditionalize a unit depending on whether an
2433 AC power source is connected or not, of whether the system
2434 is running on battery power.
2436 * systemctl gained a new "is-failed" verb that may be used in
2437 shell scripts and suchlike to check whether a specific unit
2438 is in the "failed" state.
2440 * The EnvironmentFile= setting in unit files now supports file
2441 globbing, and can hence be used to easily read a number of
2442 environment files at once.
2444 * systemd will no longer detect and recognize specific
2445 distributions. All distribution-specific #ifdeffery has been
2446 removed, systemd is now fully generic and
2447 distribution-agnostic. Effectively, not too much is lost as
2448 a lot of the code is still accessible via explicit configure
2449 switches. However, support for some distribution specific
2450 legacy configuration file formats has been dropped. We
2451 recommend distributions to simply adopt the configuration
2452 files everybody else uses now and convert the old
2453 configuration from packaging scripts. Most distributions
2454 already did that. If that's not possible or desirable,
2455 distributions are welcome to forward port the specific
2456 pieces of code locally from the git history.
2458 * When logging a message about a unit systemd will now always
2459 log the unit name in the message meta data.
2461 * localectl will now also discover system locale data that is
2462 not stored in locale archives, but directly unpacked.
2464 * logind will no longer unconditionally use framebuffer
2465 devices as seat masters, i.e. as devices that are required
2466 to be existing before a seat is considered preset. Instead,
2467 it will now look for all devices that are tagged as
2468 "seat-master" in udev. By default framebuffer devices will
2469 be marked as such, but depending on local systems other
2470 devices might be marked as well. This may be used to
2471 integrate graphics cards using closed source drivers (such
2472 as NVidia ones) more nicely into logind. Note however, that
2473 we recommend using the open source NVidia drivers instead,
2474 and no udev rules for the closed-source drivers will be
2475 shipped from us upstream.
2477 Contributions from: Adam Williamson, Alessandro Crismani, Auke
2478 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
2479 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
2480 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
2481 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
2482 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
2483 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
2484 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
2485 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
2486 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
2487 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
2488 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
2493 * udev gained support for loading additional device properties
2494 from an indexed database that is keyed by vendor/product IDs
2495 and similar device identifiers. For the beginning this
2496 "hwdb" is populated with data from the well-known PCI and
2497 USB database, but also includes PNP, ACPI and OID data. In
2498 the longer run this indexed database shall grow into
2499 becoming the one central database for non-essential
2500 userspace device metadata. Previously, data from the PCI/USB
2501 database was only attached to select devices, since the
2502 lookup was a relatively expensive operation due to O(n) time
2503 complexity (with n being the number of entries in the
2504 database). Since this is now O(1), we decided to add in this
2505 data for all devices where this is available, by
2506 default. Note that the indexed database needs to be rebuilt
2507 when new data files are installed. To achieve this you need
2508 to update your packaging scripts to invoke "udevadm hwdb
2509 --update" after installation of hwdb data files. For
2510 RPM-based distributions we introduced the new
2511 %udev_hwdb_update macro for this purpose.
2513 * The Journal gained support for the "Message Catalog", an
2514 indexed database to link up additional information with
2515 journal entries. For further details please check:
2517 http://www.freedesktop.org/wiki/Software/systemd/catalog
2519 The indexed message catalog database also needs to be
2520 rebuilt after installation of message catalog files. Use
2521 "journalctl --update-catalog" for this. For RPM-based
2522 distributions we introduced the %journal_catalog_update
2523 macro for this purpose.
2525 * The Python Journal bindings gained support for the standard
2526 Python logging framework.
2528 * The Journal API gained new functions for checking whether
2529 the underlying file system of a journal file is capable of
2530 properly reporting file change notifications, or whether
2531 applications that want to reflect journal changes "live"
2532 need to recheck journal files continuously in appropriate
2535 * It is now possible to set the "age" field for tmpfiles
2536 entries to 0, indicating that files matching this entry
2537 shall always be removed when the directories are cleaned up.
2539 * coredumpctl gained a new "gdb" verb which invokes gdb
2540 right-away on the selected coredump.
2542 * There's now support for "hybrid sleep" on kernels that
2543 support this, in addition to "suspend" and "hibernate". Use
2544 "systemctl hybrid-sleep" to make use of this.
2546 * logind's HandleSuspendKey= setting (and related settings)
2547 now gained support for a new "lock" setting to simply
2548 request the screen lock on all local sessions, instead of
2549 actually executing a suspend or hibernation.
2551 * systemd will now mount the EFI variables file system by
2554 * Socket units now gained support for configuration of the
2555 SMACK security label.
2557 * timedatectl will now output the time of the last and next
2558 daylight saving change.
2560 * We dropped support for various legacy and distro-specific
2561 concepts, such as insserv, early-boot SysV services
2562 (i.e. those for non-standard runlevels such as 'b' or 'S')
2563 or ArchLinux /etc/rc.conf support. We recommend the
2564 distributions who still need support this to either continue
2565 to maintain the necessary patches downstream, or find a
2566 different solution. (Talk to us if you have questions!)
2568 * Various systemd components will now bypass PolicyKit checks
2569 for root and otherwise handle properly if PolicyKit is not
2570 found to be around. This should fix most issues for
2571 PolicyKit-less systems. Quite frankly this should have been
2572 this way since day one. It is absolutely our intention to
2573 make systemd work fine on PolicyKit-less systems, and we
2574 consider it a bug if something does not work as it should if
2575 PolicyKit is not around.
2577 * For embedded systems it is now possible to build udev and
2578 systemd without blkid and/or kmod support.
2580 * "systemctl switch-root" is now capable of switching root
2581 more than once. I.e. in addition to transitions from the
2582 initrd to the host OS it is now possible to transition to
2583 further OS images from the host. This is useful to implement
2584 offline updating tools.
2586 * Various other additions have been made to the RPM macros
2587 shipped with systemd. Use %udev_rules_update() after
2588 installing new udev rules files. %_udevhwdbdir,
2589 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
2590 %_sysctldir are now available which resolve to the right
2591 directories for packages to place various data files in.
2593 * journalctl gained the new --full switch (in addition to
2594 --all, to disable ellipsation for long messages.
2596 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
2597 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
2598 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
2599 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
2600 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
2601 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
2602 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
2603 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
2604 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
2608 * journalctl gained new --since= and --until= switches to
2609 filter by time. It also now supports nice filtering for
2610 units via --unit=/-u.
2612 * Type=oneshot services may use ExecReload= and do the
2615 * The journal daemon now supports time-based rotation and
2616 vacuuming, in addition to the usual disk-space based
2619 * The journal will now index the available field values for
2620 each field name. This enables clients to show pretty drop
2621 downs of available match values when filtering. The bash
2622 completion of journalctl has been updated
2623 accordingly. journalctl gained a new switch -F to list all
2624 values a certain field takes in the journal database.
2626 * More service events are now written as structured messages
2627 to the journal, and made recognizable via message IDs.
2629 * The timedated, localed and hostnamed mini-services which
2630 previously only provided support for changing time, locale
2631 and hostname settings from graphical DEs such as GNOME now
2632 also have a minimal (but very useful) text-based client
2633 utility each. This is probably the nicest way to changing
2634 these settings from the command line now, especially since
2635 it lists available options and is fully integrated with bash
2638 * There's now a new tool "systemd-coredumpctl" to list and
2639 extract coredumps from the journal.
2641 * We now install a README each in /var/log/ and
2642 /etc/rc.d/init.d explaining where the system logs and init
2643 scripts went. This hopefully should help folks who go to
2644 that dirs and look into the otherwise now empty void and
2645 scratch their heads.
2647 * When user-services are invoked (by systemd --user) the
2648 $MANAGERPID env var is set to the PID of systemd.
2650 * SIGRTMIN+24 when sent to a --user instance will now result
2651 in immediate termination of systemd.
2653 * gatewayd received numerous feature additions such as a
2654 "follow" mode, for live syncing and filtering.
2656 * browse.html now allows filtering and showing detailed
2657 information on specific entries. Keyboard navigation and
2658 mouse screen support has been added.
2660 * gatewayd/journalctl now supports HTML5/JSON
2661 Server-Sent-Events as output.
2663 * The SysV init script compatibility logic will now
2664 heuristically determine whether a script supports the
2665 "reload" verb, and only then make this available as
2668 * "systemctl status --follow" has been removed, use "journalctl
2671 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
2672 have been removed since they are hardly useful to be
2675 * And I'd like to take the opportunity to specifically mention
2676 Zbigniew for his great contributions. Zbigniew, you rock!
2678 Contributions from: Andrew Eikum, Christian Hesse, Colin
2679 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
2680 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
2681 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
2682 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
2683 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
2684 Jędrzejewski-Szmek, Сковорода Никита Андреевич
2688 * If /etc/vconsole.conf is non-existent or empty we will no
2689 longer load any console font or key map at boot by
2690 default. Instead the kernel defaults will be left
2691 intact. This is definitely the right thing to do, as no
2692 configuration should mean no configuration, and hard-coding
2693 font names that are different on all archs is probably a bad
2694 idea. Also, the kernel default key map and font should be
2695 good enough for most cases anyway, and mostly identical to
2696 the userspace fonts/key maps we previously overloaded them
2697 with. If distributions want to continue to default to a
2698 non-kernel font or key map they should ship a default
2699 /etc/vconsole.conf with the appropriate contents.
2701 Contributions from: Colin Walters, Daniel J Walsh, Dave
2702 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
2703 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2707 * journalctl gained a new --cursor= switch to show entries
2708 starting from the specified location in the journal.
2710 * We now enforce a size limit on journal entry fields exported
2711 with "-o json" in journalctl. Fields larger than 4K will be
2712 assigned null. This can be turned off with --all.
2714 * An (optional) journal gateway daemon is now available as
2715 "systemd-journal-gatewayd.service". This service provides
2716 access to the journal via HTTP and JSON. This functionality
2717 will be used to implement live log synchronization in both
2718 pull and push modes, but has various other users too, such
2719 as easy log access for debugging of embedded devices. Right
2720 now it is already useful to retrieve the journal via HTTP:
2722 # systemctl start systemd-journal-gatewayd.service
2723 # wget http://localhost:19531/entries
2725 This will download the journal contents in a
2726 /var/log/messages compatible format. The same as JSON:
2728 # curl -H"Accept: application/json" http://localhost:19531/entries
2730 This service is also accessible via a web browser where a
2731 single static HTML5 app is served that uses the JSON logic
2732 to enable the user to do some basic browsing of the
2733 journal. This will be extended later on. Here's an example
2734 screenshot of this app in its current state:
2736 http://0pointer.de/public/journal-gatewayd
2738 Contributions from: Kay Sievers, Lennart Poettering, Robert
2739 Milasan, Tom Gundersen
2743 * The bash completion logic is now available for journalctl
2746 * We do not mount the "cpuset" controller anymore together with
2747 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
2748 started if no parameters are assigned to it. "cpuset" hence
2749 broke code that assumed it it could create "cpu" groups and
2752 * journalctl -f will now subscribe to terminal size changes,
2753 and line break accordingly.
2755 Contributions from: Dave Reisner, Kay Sievers, Lennart
2756 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
2760 * nspawn will now create a symlink /etc/localtime in the
2761 container environment, copying the host's timezone
2762 setting. Previously this has been done via a bind mount, but
2763 since symlinks cannot be bind mounted this has now been
2764 changed to create/update the appropriate symlink.
2766 * journalctl -n's line number argument is now optional, and
2767 will default to 10 if omitted.
2769 * journald will now log the maximum size the journal files may
2770 take up on disk. This is particularly useful if the default
2771 built-in logic of determining this parameter from the file
2772 system size is used. Use "systemctl status
2773 systemd-journald.service" to see this information.
2775 * The multi-seat X wrapper tool has been stripped down. As X
2776 is now capable of enumerating graphics devices via udev in a
2777 seat-aware way the wrapper is not strictly necessary
2778 anymore. A stripped down temporary stop-gap is still shipped
2779 until the upstream display managers have been updated to
2780 fully support the new X logic. Expect this wrapper to be
2781 removed entirely in one of the next releases.
2783 * HandleSleepKey= in logind.conf has been split up into
2784 HandleSuspendKey= and HandleHibernateKey=. The old setting
2785 is not available anymore. X11 and the kernel are
2786 distuingishing between these keys and we should too. This
2787 also means the inhibition lock for these keys has been split
2790 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
2791 Poettering, Lukas Nykryn, Václav Pavlín
2795 * Whenever a unit changes state we will now log this to the
2796 journal and show along the unit's own log output in
2799 * ConditionPathIsMountPoint= can now properly detect bind
2800 mount points too. (Previously, a bind mount of one file
2801 system to another place in the same file system could not be
2802 detected as mount, since they shared struct stat's st_dev
2805 * We will now mount the cgroup controllers cpu, cpuacct,
2806 cpuset and the controllers net_cls, net_prio together by
2809 * nspawn containers will now have a virtualized boot
2810 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
2811 over with a randomized ID at container initialization). This
2812 has the effect of making "journalctl -b" do the right thing
2815 * The JSON output journal serialization has been updated not
2816 to generate "endless" list objects anymore, but rather one
2817 JSON object per line. This is more in line how most JSON
2818 parsers expect JSON objects. The new output mode
2819 "json-pretty" has been added to provide similar output, but
2820 neatly aligned for readability by humans.
2822 * We dropped all explicit sync() invocations in the shutdown
2823 code. The kernel does this implicitly anyway in the kernel
2824 reboot() syscall. halt(8)'s -n option is now a compatibility
2827 * We now support virtualized reboot() in containers, as
2828 supported by newer kernels. We will fall back to exit() if
2829 CAP_SYS_REBOOT is not available to the container. Also,
2830 nspawn makes use of this now and will actually reboot the
2831 container if the containerized OS asks for that.
2833 * journalctl will only show local log output by default
2834 now. Use --merge (-m) to show remote log output, too.
2836 * libsystemd-journal gained the new sd_journal_get_usage()
2837 call to determine the current disk usage of all journal
2838 files. This is exposed in the new "journalctl --disk-usage"
2841 * journald gained a new configuration setting SplitMode= in
2842 journald.conf which may be used to control how user journals
2843 are split off. See journald.conf(5) for details.
2845 * A new condition type ConditionFileNotEmpty= has been added.
2847 * tmpfiles' "w" lines now support file globbing, to write
2848 multiple files at once.
2850 * We added Python bindings for the journal submission
2851 APIs. More Python APIs for a number of selected APIs will
2852 likely follow. Note that we intend to add native bindings
2853 only for the Python language, as we consider it common
2854 enough to deserve bindings shipped within systemd. There are
2855 various projects outside of systemd that provide bindings
2856 for languages such as PHP or Lua.
2858 * Many conditions will now resolve specifiers such as %i. In
2859 addition, PathChanged= and related directives of .path units
2860 now support specifiers as well.
2862 * There's now a new RPM macro definition for the system preset
2865 * journald will now warn if it ca not forward a message to the
2866 syslog daemon because its socket is full.
2868 * timedated will no longer write or process /etc/timezone,
2869 except on Debian. As we do not support late mounted /usr
2870 anymore /etc/localtime always being a symlink is now safe,
2871 and hence the information in /etc/timezone is not necessary
2874 * logind will now always reserve one VT for a text getty (VT6
2875 by default). Previously if more than 6 X sessions where
2876 started they took up all the VTs with auto-spawned gettys,
2877 so that no text gettys were available anymore.
2879 * udev will now automatically inform the btrfs kernel logic
2880 about btrfs RAID components showing up. This should make
2881 simple hotplug based btrfs RAID assembly work.
2883 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
2884 (but not for its children which will stay at the kernel
2885 default). This should allow setups with a lot more listening
2888 * systemd will now always pass the configured timezone to the
2889 kernel at boot. timedated will do the same when the timezone
2892 * logind's inhibition logic has been updated. By default,
2893 logind will now handle the lid switch, the power and sleep
2894 keys all the time, even in graphical sessions. If DEs want
2895 to handle these events on their own they should take the new
2896 handle-power-key, handle-sleep-key and handle-lid-switch
2897 inhibitors during their runtime. A simple way to achiveve
2898 that is to invoke the DE wrapped in an invocation of:
2900 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
2902 * Access to unit operations is now checked via SELinux taking
2903 the unit file label and client process label into account.
2905 * systemd will now notify the administrator in the journal
2906 when he over-mounts a non-empty directory.
2908 * There are new specifiers that are resolved in unit files,
2909 for the host name (%H), the machine ID (%m) and the boot ID
2912 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
2913 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
2914 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
2915 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
2916 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
2917 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2918 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
2922 * Support for reading structured kernel messages from
2923 /dev/kmsg has now been added and is enabled by default.
2925 * Support for reading kernel messages from /proc/kmsg has now
2926 been removed. If you want kernel messages in the journal
2927 make sure to run a recent kernel (>= 3.5) that supports
2928 reading structured messages from /dev/kmsg (see
2929 above). /proc/kmsg is now exclusive property of classic
2930 syslog daemons again.
2932 * The libudev API gained the new
2933 udev_device_new_from_device_id() call.
2935 * The logic for file system namespace (ReadOnlyDirectory=,
2936 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
2937 require pivot_root() anymore. This means fewer temporary
2938 directories are created below /tmp for this feature.
2940 * nspawn containers will now see and receive all submounts
2941 made on the host OS below the root file system of the
2944 * Forward Secure Sealing is now supported for Journal files,
2945 which provide cryptographical sealing of journal files so
2946 that attackers cannot alter log history anymore without this
2947 being detectable. Lennart will soon post a blog story about
2948 this explaining it in more detail.
2950 * There are two new service settings RestartPreventExitStatus=
2951 and SuccessExitStatus= which allow configuration of exit
2952 status (exit code or signal) which will be excepted from the
2953 restart logic, resp. consider successful.
2955 * journalctl gained the new --verify switch that can be used
2956 to check the integrity of the structure of journal files and
2957 (if Forward Secure Sealing is enabled) the contents of
2960 * nspawn containers will now be run with /dev/stdin, /dev/fd/
2961 and similar symlinks pre-created. This makes running shells
2962 as container init process a lot more fun.
2964 * The fstab support can now handle PARTUUID= and PARTLABEL=
2967 * A new ConditionHost= condition has been added to match
2968 against the hostname (with globs) and machine ID. This is
2969 useful for clusters where a single OS image is used to
2970 provision a large number of hosts which shall run slightly
2971 different sets of services.
2973 * Services which hit the restart limit will now be placed in a
2976 Contributions from: Bertram Poettering, Dave Reisner, Huang
2977 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
2978 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
2982 * When running in --user mode systemd will now become a
2983 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
2984 tree a lot more organized.
2986 * A new PartOf= unit dependency type has been introduced that
2987 may be used to group services in a natural way.
2989 * "systemctl enable" may now be used to enable instances of
2992 * journalctl now prints error log levels in red, and
2993 warning/notice log levels in bright white. It also supports
2994 filtering by log level now.
2996 * cgtop gained a new -n switch (similar to top), to configure
2997 the maximum number of iterations to run for. It also gained
2998 -b, to run in batch mode (accepting no input).
3000 * The suffix ".service" may now be omitted on most systemctl
3001 command lines involving service unit names.
3003 * There's a new bus call in logind to lock all sessions, as
3004 well as a loginctl verb for it "lock-sessions".
3006 * libsystemd-logind.so gained a new call sd_journal_perror()
3007 that works similar to libc perror() but logs to the journal
3008 and encodes structured information about the error number.
3010 * /etc/crypttab entries now understand the new keyfile-size=
3013 * shutdown(8) now can send a (configurable) wall message when
3014 a shutdown is cancelled.
3016 * The mount propagation mode for the root file system will now
3017 default to "shared", which is useful to make containers work
3018 nicely out-of-the-box so that they receive new mounts from
3019 the host. This can be undone locally by running "mount
3020 --make-rprivate /" if needed.
3022 * The prefdm.service file has been removed. Distributions
3023 should maintain this unit downstream if they intend to keep
3024 it around. However, we recommend writing normal unit files
3025 for display managers instead.
3027 * Since systemd is a crucial part of the OS we will now
3028 default to a number of compiler switches that improve
3029 security (hardening) such as read-only relocations, stack
3030 protection, and suchlike.
3032 * The TimeoutSec= setting for services is now split into
3033 TimeoutStartSec= and TimeoutStopSec= to allow configuration
3034 of individual time outs for the start and the stop phase of
3037 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
3038 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
3039 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
3040 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
3041 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
3042 Gundersen, Zbigniew Jędrzejewski-Szmek
3046 * The journal and id128 C APIs are now fully documented as man
3049 * Extra safety checks have been added when transitioning from
3050 the initial RAM disk to the main system to avoid accidental
3053 * /etc/crypttab entries now understand the new keyfile-offset=
3056 * systemctl -t can now be used to filter by unit load state.
3058 * The journal C API gained the new sd_journal_wait() call to
3059 make writing synchronous journal clients easier.
3061 * journalctl gained the new -D switch to show journals from a
3064 * journalctl now displays a special marker between log
3065 messages of two different boots.
3067 * The journal is now explicitly flushed to /var via a service
3068 systemd-journal-flush.service, rather than implicitly simply
3069 by seeing /var/log/journal to be writable.
3071 * journalctl (and the journal C APIs) can now match for much
3072 more complex expressions, with alternatives and
3075 * When transitioning from the initial RAM disk to the main
3076 system we will now kill all processes in a killing spree to
3077 ensure no processes stay around by accident.
3079 * Three new specifiers may be used in unit files: %u, %h, %s
3080 resolve to the user name, user home directory resp. user
3081 shell. This is useful for running systemd user instances.
3083 * We now automatically rotate journal files if their data
3084 object hash table gets a fill level > 75%. We also size the
3085 hash table based on the configured maximum file size. This
3086 together should lower hash collisions drastically and thus
3087 speed things up a bit.
3089 * journalctl gained the new "--header" switch to introspect
3090 header data of journal files.
3092 * A new setting SystemCallFilters= has been added to services
3093 which may be used to apply blacklists or whitelists to
3094 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
3096 * nspawn gained a new --link-journal= switch (and quicker: -j)
3097 to link the container journal with the host. This makes it
3098 very easy to centralize log viewing on the host for all
3099 guests while still keeping the journal files separated.
3101 * Many bugfixes and optimizations
3103 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
3104 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
3105 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
3110 * Several tools now understand kernel command line arguments,
3111 which are only read when run in an initial RAM disk. They
3112 usually follow closely their normal counterparts, but are
3115 * There's a new tool to analyze the readahead files that are
3116 automatically generated at boot. Use:
3118 /usr/lib/systemd/systemd-readahead analyze /.readahead
3120 * We now provide an early debug shell on tty9 if this enabled. Use:
3122 systemctl enable debug-shell.service
3124 * All plymouth related units have been moved into the Plymouth
3125 package. Please make sure to upgrade your Plymouth version
3128 * systemd-tmpfiles now supports getting passed the basename of
3129 a configuration file only, in which case it will look for it
3130 in all appropriate directories automatically.
3132 * udevadm info now takes a /dev or /sys path as argument, and
3133 does the right thing. Example:
3135 udevadm info /dev/sda
3136 udevadm info /sys/class/block/sda
3138 * systemctl now prints a warning if a unit is stopped but a
3139 unit that might trigger it continues to run. Example: a
3140 service is stopped but the socket that activates it is left
3143 * "systemctl status" will now mention if the log output was
3144 shortened due to rotation since a service has been started.
3146 * The journal API now exposes functions to determine the
3147 "cutoff" times due to rotation.
3149 * journald now understands SIGUSR1 and SIGUSR2 for triggering
3150 immediately flushing of runtime logs to /var if possible,
3151 resp. for triggering immediate rotation of the journal
3154 * It is now considered an error if a service is attempted to
3155 be stopped that is not loaded.
3157 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
3159 * systemd-analyze now supports Python 3
3161 * tmpfiles now supports cleaning up directories via aging
3162 where the first level dirs are always kept around but
3163 directories beneath it automatically aged. This is enabled
3164 by prefixing the age field with '~'.
3166 * Seat objects now expose CanGraphical, CanTTY properties
3167 which is required to deal with very fast bootups where the
3168 display manager might be running before the graphics drivers
3169 completed initialization.
3171 * Seat objects now expose a State property.
3173 * We now include RPM macros for service enabling/disabling
3174 based on the preset logic. We recommend RPM based
3175 distributions to make use of these macros if possible. This
3176 makes it simpler to reuse RPM spec files across
3179 * We now make sure that the collected systemd unit name is
3180 always valid when services log to the journal via
3183 * There's a new man page kernel-command-line(7) detailing all
3184 command line options we understand.
3186 * The fstab generator may now be disabled at boot by passing
3187 fstab=0 on the kernel command line.
3189 * A new kernel command line option modules-load= is now understood
3190 to load a specific kernel module statically, early at boot.
3192 * Unit names specified on the systemctl command line are now
3193 automatically escaped as needed. Also, if file system or
3194 device paths are specified they are automatically turned
3195 into the appropriate mount or device unit names. Example:
3197 systemctl status /home
3198 systemctl status /dev/sda
3200 * The SysVConsole= configuration option has been removed from
3201 system.conf parsing.
3203 * The SysV search path is no longer exported on the D-Bus
3206 * The Names= option is been removed from unit file parsing.
3208 * There's a new man page bootup(7) detailing the boot process.
3210 * Every unit and every generator we ship with systemd now
3211 comes with full documentation. The self-explanatory boot is
3214 * A couple of services gained "systemd-" prefixes in their
3215 name if they wrap systemd code, rather than only external
3216 code. Among them fsck@.service which is now
3217 systemd-fsck@.service.
3219 * The HaveWatchdog property has been removed from the D-Bus
3222 * systemd.confirm_spawn= on the kernel command line should now
3225 * There's a new man page crypttab(5) which details all options
3226 we actually understand.
3228 * systemd-nspawn gained a new --capability= switch to pass
3229 additional capabilities to the container.
3231 * timedated will now read known NTP implementation unit names
3232 from /usr/lib/systemd/ntp-units.d/*.list,
3233 systemd-timedated-ntp.target has been removed.
3235 * journalctl gained a new switch "-b" that lists log data of
3236 the current boot only.
3238 * The notify socket is in the abstract namespace again, in
3239 order to support daemons which chroot() at start-up.
3241 * There is a new Storage= configuration option for journald
3242 which allows configuration of where log data should go. This
3243 also provides a way to disable journal logging entirely, so
3244 that data collected is only forwarded to the console, the
3245 kernel log buffer or another syslog implementation.
3247 * Many bugfixes and optimizations
3249 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
3250 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
3251 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
3252 Shawn Landden, Tom Gundersen
3256 * "systemctl help <unit>" now shows the man page if one is
3259 * Several new man pages have been added.
3261 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
3262 MaxLevelConsole= can now be specified in
3263 journald.conf. These options allow reducing the amount of
3264 data stored on disk or forwarded by the log level.
3266 * TimerSlackNSec= can now be specified in system.conf for
3267 PID1. This allows system-wide power savings.
3269 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
3270 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
3275 * logind is now capable of (optionally) handling power and
3276 sleep keys as well as the lid switch.
3278 * journalctl now understands the syntax "journalctl
3279 /usr/bin/avahi-daemon" to get all log output of a specific
3282 * CapabilityBoundingSet= in system.conf now also influences
3283 the capability bound set of usermode helpers of the kernel.
3285 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
3286 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
3287 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
3288 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
3292 * Note that we skipped 139 releases here in order to set the
3293 new version to something that is greater than both udev's
3294 and systemd's most recent version number.
3296 * udev: all udev sources are merged into the systemd source tree now.
3297 All future udev development will happen in the systemd tree. It
3298 is still fully supported to use the udev daemon and tools without
3299 systemd running, like in initramfs or other init systems. Building
3300 udev though, will require the *build* of the systemd tree, but
3301 udev can be properly *run* without systemd.
3303 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
3304 should be used to create dead device nodes as workarounds for broken
3307 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
3308 no longer supported. udev_monitor_new_from_netlink() needs to be
3309 used to subscribe to events.
3311 * udev: when udevd is started by systemd, processes which are left
3312 behind by forking them off of udev rules, are unconditionally cleaned
3313 up and killed now after the event handling has finished. Services or
3314 daemons must be started as systemd services. Services can be
3315 pulled-in by udev to get started, but they can no longer be directly
3316 forked by udev rules.
3318 * udev: the daemon binary is called systemd-udevd now and installed
3319 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
3320 to adapt to that, create symlink, or rename the binary after building
3323 * libudev no longer provides these symbols:
3324 udev_monitor_from_socket()
3325 udev_queue_get_failed_list_entry()
3326 udev_get_{dev,sys,run}_path()
3327 The versions number was bumped and symbol versioning introduced.
3329 * systemd-loginctl and systemd-journalctl have been renamed
3330 to loginctl and journalctl to match systemctl.
3332 * The config files: /etc/systemd/systemd-logind.conf and
3333 /etc/systemd/systemd-journald.conf have been renamed to
3334 logind.conf and journald.conf. Package updates should rename
3335 the files to the new names on upgrade.
3337 * For almost all files the license is now LGPL2.1+, changed
3338 from the previous GPL2.0+. Exceptions are some minor stuff
3339 of udev (which will be changed to LGPL2.1 eventually, too),
3340 and the MIT licensed sd-daemon.[ch] library that is suitable
3341 to be used as drop-in files.
3343 * systemd and logind now handle system sleep states, in
3344 particular suspending and hibernating.
3346 * logind now implements a sleep/shutdown/idle inhibiting logic
3347 suitable for a variety of uses. Soonishly Lennart will blog
3348 about this in more detail.
3350 * var-run.mount and var-lock.mount are no longer provided
3351 (which prevously bind mounted these directories to their new
3352 places). Distributions which have not converted these
3353 directories to symlinks should consider stealing these files
3354 from git history and add them downstream.
3356 * We introduced the Documentation= field for units and added
3357 this to all our shipped units. This is useful to make it
3358 easier to explore the boot and the purpose of the various
3361 * All smaller setup units (such as
3362 systemd-vconsole-setup.service) now detect properly if they
3363 are run in a container and are skipped when
3364 appropriate. This guarantees an entirely noise-free boot in
3365 Linux container environments such as systemd-nspawn.
3367 * A framework for implementing offline system updates is now
3368 integrated, for details see:
3369 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
3371 * A new service type Type=idle is available now which helps us
3372 avoiding ugly interleaving of getty output and boot status
3375 * There's now a system-wide CapabilityBoundingSet= option to
3376 globally reduce the set of capabilities for the
3377 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
3378 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
3379 even CAP_NET_ADMIN system-wide for secure systems.
3381 * There are now system-wide DefaultLimitXXX= options to
3382 globally change the defaults of the various resource limits
3383 for all units started by PID 1.
3385 * Harald Hoyer's systemd test suite has been integrated into
3386 systemd which allows easy testing of systemd builds in qemu
3387 and nspawn. (This is really awesome! Ask us for details!)
3389 * The fstab parser is now implemented as generator, not inside
3392 * systemctl will now warn you if .mount units generated from
3393 /etc/fstab are out of date due to changes in fstab that
3394 have not been read by systemd yet.
3396 * systemd is now suitable for usage in initrds. Dracut has
3397 already been updated to make use of this. With this in place
3398 initrds get a slight bit faster but primarily are much
3399 easier to introspect and debug since "systemctl status" in
3400 the host system can be used to introspect initrd services,
3401 and the journal from the initrd is kept around too.
3403 * systemd-delta has been added, a tool to explore differences
3404 between user/admin configuration and vendor defaults.
3406 * PrivateTmp= now affects both /tmp and /var/tmp.
3408 * Boot time status messages are now much prettier and feature
3409 proper english language. Booting up systemd has never been
3412 * Read-ahead pack files now include the inode number of all
3413 files to pre-cache. When the inode changes the pre-caching
3414 is not attempted. This should be nicer to deal with updated
3415 packages which might result in changes of read-ahead
3418 * We now temporaritly lower the kernel's read_ahead_kb variable
3419 when collecting read-ahead data to ensure the kernel's
3420 built-in read-ahead does not add noise to our measurements
3421 of necessary blocks to pre-cache.
3423 * There's now RequiresMountsFor= to add automatic dependencies
3424 for all mounts necessary for a specific file system path.
3426 * MountAuto= and SwapAuto= have been removed from
3427 system.conf. Mounting file systems at boot has to take place
3430 * nspawn now learned a new switch --uuid= to set the machine
3431 ID on the command line.
3433 * nspawn now learned the -b switch to automatically search
3436 * vt102 is now the default TERM for serial TTYs, upgraded from
3439 * systemd-logind now works on VT-less systems.
3441 * The build tree has been reorganized. The individual
3442 components now have directories of their own.
3444 * A new condition type ConditionPathIsReadWrite= is now available.
3446 * nspawn learned the new -C switch to create cgroups for the
3447 container in other hierarchies.
3449 * We now have support for hardware watchdogs, configurable in
3452 * The scheduled shutdown logic now has a public API.
3454 * We now mount /tmp as tmpfs by default, but this can be
3455 masked and /etc/fstab can override it.
3457 * Since udisks does not make use of /media anymore we are not
3458 mounting a tmpfs on it anymore.
3460 * journalctl gained a new --local switch to only interleave
3461 locally generated journal files.
3463 * We can now load the IMA policy at boot automatically.
3465 * The GTK tools have been split off into a systemd-ui.
3467 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
3468 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
3469 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
3470 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
3471 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
3472 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
3473 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
3474 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
3475 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
3480 * This is mostly a bugfix release
3482 * Support optional initialization of the machine ID from the
3483 KVM or container configured UUID.
3485 * Support immediate reboots with "systemctl reboot -ff"
3487 * Show /etc/os-release data in systemd-analyze output
3489 * Many bugfixes for the journal, including endianness fixes and
3490 ensuring that disk space enforcement works
3492 * sd-login.h is C++ comptaible again
3494 * Extend the /etc/os-release format on request of the Debian
3497 * We now refuse non-UTF8 strings used in various configuration
3498 and unit files. This is done to ensure we do not pass invalid
3499 data over D-Bus or expose it elsewhere.
3501 * Register Mimo USB Screens as suitable for automatic seat
3504 * Read SELinux client context from journal clients in a race
3507 * Reorder configuration file lookup order. /etc now always
3508 overrides /run in order to allow the administrator to always
3509 and unconditionally override vendor supplied or
3510 automatically generated data.
3512 * The various user visible bits of the journal now have man
3513 pages. We still lack man pages for the journal API calls
3516 * We now ship all man pages in HTML format again in the
3519 Contributions from: Dave Reisner, Dirk Eibach, Frederic
3520 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
3521 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
3526 * This is mostly a bugfix release
3528 * systems lacking /etc/os-release are no longer supported.
3530 * Various functionality updates to libsystemd-login.so
3532 * Track class of PAM logins to distuingish greeters from
3535 Contributions from: Kay Sievers, Lennart Poettering, Michael
3540 * This is an important bugfix release for v41.
3542 * Building man pages is now optional which should be useful
3543 for those building systemd from git but unwilling to install
3546 * Watchdog support for supervising services is now usable. In
3547 a future release support for hardware watchdogs
3548 (i.e. /dev/watchdog) will be added building on this.
3550 * Service start rate limiting is now configurable and can be
3551 turned off per service. When a start rate limit is hit a
3552 reboot can automatically be triggered.
3554 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
3556 Contributions from: Benjamin Franzke, Bill Nottingham,
3557 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
3558 Schmidt, Michał Górny, Piotr Drąg
3562 * The systemd binary is installed /usr/lib/systemd/systemd now;
3563 An existing /sbin/init symlink needs to be adapted with the
3566 * The code that loads kernel modules has been ported to invoke
3567 libkmod directly, instead of modprobe. This means we do not
3568 support systems with module-init-tools anymore.
3570 * Watchdog support is now already useful, but still not
3573 * A new kernel command line option systemd.setenv= is
3574 understood to set system wide environment variables
3575 dynamically at boot.
3577 * We now limit the set of capabilities of systemd-journald.
3579 * We now set SIGPIPE to ignore by default, since it only is
3580 useful in shell pipelines, and has little use in general
3581 code. This can be disabled with IgnoreSIPIPE=no in unit
3584 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
3585 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
3590 * This is mostly a bugfix release
3592 * We now expose the reason why a service failed in the
3593 "Result" D-Bus property.
3595 * Rudimentary service watchdog support (will be completed over
3596 the next few releases.)
3598 * When systemd forks off in order execute some service we will
3599 now immediately changes its argv[0] to reflect which process
3600 it will execute. This is useful to minimize the time window
3601 with a generic argv[0], which makes bootcharts more useful
3603 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
3604 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
3605 Mike Kazantsev, Ray Strode
3609 * This is mostly a test release, but incorporates many
3612 * New systemd-cgtop tool to show control groups by their
3615 * Linking against libacl for ACLs is optional again. If
3616 disabled, support tracking device access for active logins
3617 goes becomes unavailable, and so does access to the user
3618 journals by the respective users.
3620 * If a group "adm" exists, journal files are automatically
3621 owned by them, thus allow members of this group full access
3622 to the system journal as well as all user journals.
3624 * The journal now stores the SELinux context of the logging
3625 client for all entries.
3627 * Add C++ inclusion guards to all public headers
3629 * New output mode "cat" in the journal to print only text
3630 messages, without any meta data like date or time.
3632 * Include tiny X server wrapper as a temporary stop-gap to
3633 teach XOrg udev display enumeration. This is used by display
3634 managers such as gdm, and will go away as soon as XOrg
3635 learned native udev hotplugging for display devices.
3637 * Add new systemd-cat tool for executing arbitrary programs
3638 with STDERR/STDOUT connected to the journal. Can also act as
3639 BSD logger replacement, and does so by default.
3641 * Optionally store all locally generated coredumps in the
3642 journal along with meta data.
3644 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
3645 writing short strings to files (for usage for /sys), and for
3646 creating symlinks, character and block device nodes.
3648 * New unit file option ControlGroupPersistent= to make cgroups
3649 persistent, following the mechanisms outlined in
3650 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
3652 * Support multiple local RTCs in a sane way
3654 * No longer monopolize IO when replaying readahead data on
3655 rotating disks, since we might starve non-file-system IO to
3656 death, since fanotify() will not see accesses done by blkid,
3659 * Do not show kernel threads in systemd-cgls anymore, unless
3660 requested with new -k switch.
3662 Contributions from: Dan Horák, Kay Sievers, Lennart
3663 Poettering, Michal Schmidt
3667 * This is mostly a test release, but incorporates many
3670 * The git repository moved to:
3671 git://anongit.freedesktop.org/systemd/systemd
3672 ssh://git.freedesktop.org/git/systemd/systemd
3674 * First release with the journal
3675 http://0pointer.de/blog/projects/the-journal.html
3677 * The journal replaces both systemd-kmsg-syslogd and
3678 systemd-stdout-bridge.
3680 * New sd_pid_get_unit() API call in libsystemd-logind
3682 * Many systemadm clean-ups
3684 * Introduce remote-fs-pre.target which is ordered before all
3685 remote mounts and may be used to start services before all
3688 * Added Mageia support
3690 * Add bash completion for systemd-loginctl
3692 * Actively monitor PID file creation for daemons which exit in
3693 the parent process before having finished writing the PID
3694 file in the daemon process. Daemons which do this need to be
3695 fixed (i.e. PID file creation must have finished before the
3696 parent exits), but we now react a bit more gracefully to them.
3698 * Add colourful boot output, mimicking the well-known output
3699 of existing distributions.
3701 * New option PassCredentials= for socket units, for
3702 compatibility with a recent kernel ABI breakage.
3704 * /etc/rc.local is now hooked in via a generator binary, and
3705 thus will no longer act as synchronization point during
3708 * systemctl list-unit-files now supports --root=.
3710 * systemd-tmpfiles now understands two new commands: z, Z for
3711 relabelling files according to the SELinux database. This is
3712 useful to apply SELinux labels to specific files in /sys,
3715 * Output of SysV services is now forwarded to both the console
3716 and the journal by default, not only just the console.
3718 * New man pages for all APIs from libsystemd-login.
3720 * The build tree got reorganized and a the build system is a
3721 lot more modular allowing embedded setups to specifically
3722 select the components of systemd they are interested in.
3724 * Support for Linux systems lacking the kernel VT subsystem is
3727 * configure's --with-rootdir= got renamed to
3728 --with-rootprefix= to follow the naming used by udev and
3731 * Unless specified otherwise we will now install to /usr instead
3732 of /usr/local by default.
3734 * Processes with '@' in argv[0][0] are now excluded from the
3735 final shut-down killing spree, following the logic explained
3737 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
3739 * All processes remaining in a service cgroup when we enter
3740 the START or START_PRE states are now killed with
3741 SIGKILL. That means it is no longer possible to spawn
3742 background processes from ExecStart= lines (which was never
3743 supported anyway, and bad style).
3745 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
3746 reloading of units together.
3748 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
3749 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
3750 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
3751 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
3752 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek