1 systemd System and Service Manager
5 * A new component "systemd-networkd" has been added that can
6 be used to configure local network interfaces statically or
7 via DHCP. It is capable of bringing up bridges, VLANs, and
8 bonding. Currently, no hook-ups for interactive network
9 configuration are provided. Use this for your initrd,
10 container, embedded, or server setup if you need a simple,
11 yet powerful, network configuration solution. This
12 configuration subsystem is quite nifty, as it allows wildcard
13 hotplug matching in interfaces. For example, with a single
14 configuration snippet, you can configure that all Ethernet
15 interfaces showing up are automatically added to a bridge,
16 or similar. It supports link-sensing and more.
18 * A new tool "systemd-socket-proxyd" has been added which can
19 act as a bidirectional proxy for TCP sockets. This is
20 useful for adding socket activation support to services that
21 do not actually support socket activation, including virtual
22 machines and the like.
24 * Add a new tool to save/restore rfkill state on
27 * Save/restore state of keyboard backlights in addition to
28 display backlights on shutdown/boot.
30 * udev learned a new SECLABEL{} construct to label device
31 nodes with a specific security label when they appear. For
32 now, only SECLABEL{selinux} is supported, but the syntax is
33 prepared for additional security frameworks.
35 * udev gained a new scheme to configure link-level attributes
36 from files in /etc/systemd/network/*.link. These files can
37 match against MAC address, device path, driver name and type,
38 and will apply attributes like the naming policy, link speed,
39 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
40 address assignment policy (randomized, ...).
42 * When the User= switch is used in a unit file, also
43 initialize $SHELL= based on the user database entry.
45 * systemd no longer depends on libdbus. All communication is
46 now done with sd-bus, systemd's low-level bus library
49 * kdbus support has been added to PID 1 itself. When kdbus is
50 enabled, this causes PID 1 to set up the system bus and
51 enable support for a new ".busname" unit type that
52 encapsulates bus name activation on kdbus. It works a little
53 bit like ".socket" units, except for bus names. A new
54 generator has been added that converts classic dbus1 service
55 activation files automatically into native systemd .busname
58 * sd-bus: add a light-weight vtable implementation that allows
59 defining objects on the bus with a simple static const
60 vtable array of its methods, signals and properties.
62 * systemd will not generate or install static dbus
63 introspection data anymore to /usr/share/dbus-1/interfaces,
64 as the precise format of these files is unclear, and
65 nothing makes use of it.
67 * A proxy daemon is now provided to proxy clients connecting
68 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
69 compatibility with classic D-Bus.
71 * A bus driver implementation has been added that supports the
72 classic D-Bus bus driver calls on kdbus, also for
73 compatibility purposes.
75 * A new API "sd-event.h" has been added that implements a
76 minimal event loop API built around epoll. It provides a
77 couple of features that direct epoll usage is lacking:
78 prioritization of events, scales to large numbers of timer
79 events, per-event timer slack (accuracy), system-wide
80 coalescing of timer events, exit handlers, watchdog
81 supervision support using systemd's sd_notify() API, child
84 * A new API "sd-rntl.h" has been added that provides an API
85 around the route netlink interface of the kernel, similar in
88 * A new API "sd-dhcp-client.h" has been added that provides a
89 small DHCPv4 client-side implementation. This is used by
92 * There is a new kernel command line option
93 "systemd.restore_state=0|1". When set to "0", none of the
94 systemd tools will restore saved runtime state to hardware
95 devices. More specifically, the rfkill and backlight states
98 * The FsckPassNo= compatibility option in mount/service units
99 has been removed. The fstab generator will now add the
100 necessary dependencies automatically, and does not require
101 PID1's support for that anymore.
103 * journalctl gained a new switch, --list-boots, that lists
104 recent boots with their times and boot IDs.
106 * The various tools like systemctl, loginctl, timedatectl,
107 busctl, systemd-run, ... have gained a new switch "-M" to
108 connect to a specific, local OS container (as direct
109 connection, without requiring SSH). This works on any
110 container that is registered with machined, such as those
111 created by libvirt-lxc or nspawn.
113 * systemd-run and systemd-analyze also gained support for "-H"
114 to connect to remote hosts via SSH. This is particularly
115 useful for systemd-run because it enables queuing of jobs
118 * machinectl gained a new command "login" to open a getty
119 login in any local container. This works with any container
120 that is registered with machined (such as those created by
121 libvirt-lxc or nspawn), and which runs systemd inside.
123 * machinectl gained a new "reboot" command that may be used to
124 trigger a reboot on a specific container that is registered
125 with machined. This works on any container that runs an init
128 * systemctl gained a new "list-timers" command to print a nice
129 listing of installed timer units with the times they elapse
132 * Alternative reboot() parameters may now be specified on the
133 "systemctl reboot" command line and are passed to the
134 reboot() system call.
136 * systemctl gained a new --job-mode= switch to configure the
137 mode to queue a job with. This is a more generic version of
138 --fail, --irreversible, and --ignore-dependencies, which are
139 still available but not advertised anymore.
141 * /etc/systemd/system.conf gained new settings to configure
142 various default timeouts of units, as well as the default
143 start limit interval and burst. These may still be overridden
146 * PID1 will now export on the bus profile data of the security
147 policy upload process (such as the SELinux policy upload to
150 * journald: when forwarding logs to the console, include
151 timestamps (following the setting in
152 /sys/module/printk/parameters/time).
154 * OnCalendar= in timer units now understands the special
155 strings "yearly" and "annually". (Both are equivalent)
157 * The accuracy of timer units is now configurable with the new
158 AccuracySec= setting. It defaults to 1min.
160 * A new dependency type JoinsNamespaceOf= has been added that
161 allows running two services within the same /tmp and network
162 namespace, if PrivateNetwork= or PrivateTmp= are used.
164 * A new command "cat" has been added to systemctl. It outputs
165 the original unit file of a unit, and concatenates the
166 contents of additional "drop-in" unit file snippets, so that
167 the full configuration is shown.
169 * systemctl now supports globbing on the various "list-xyz"
170 commands, like "list-units" or "list-sockets", as well as on
171 those commands which take multiple unit names.
173 * journalctl's --unit= switch gained support for globbing.
175 * All systemd daemons now make use of the watchdog logic so
176 that systemd automatically notices when they hang.
178 * If the $container_ttys environment variable is set,
179 getty-generator will automatically spawn a getty for each
180 listed tty. This is useful for container managers to request
181 login gettys to be spawned on as many ttys as needed.
183 * %h, %s, %U specifier support is not available anymore when
184 used in unit files for PID 1. This is because NSS calls are
185 not safe from PID 1. They stay available for --user
186 instances of systemd, and as special case for the root user.
188 * loginctl gained a new "--no-legend" switch to turn off output
191 * The "sd-login.h" API gained three new calls:
192 sd_session_is_remote(), sd_session_get_remote_user(),
193 sd_session_get_remote_host() to query information about
196 * The udev hardware database now also carries vendor/product
197 information of SDIO devices.
199 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
200 determine whether watchdog notifications are requested by
203 * Socket-activated per-connection services now include a
204 short description of the connection parameters in the
207 * tmpfiles gained a new "--boot" option. When this is not used,
208 only lines where the command character is not suffixed with
209 "!" are executed. When this option is specified, those
210 options are executed too. This partitions tmpfiles
211 directives into those that can be safely executed at any
212 time, and those which should be run only at boot (for
213 example, a line that creates /run/nologin).
215 * A new API "sd-resolve.h" has been added which provides a simple
216 asynchronous wrapper around glibc NSS host name resolution
217 calls, such as getaddrinfo(). In contrast to glibc's
218 getaddrinfo_a(), it does not use signals. In contrast to most
219 other asynchronous name resolution libraries, this one does
220 not reimplement DNS, but reuses NSS, so that alternate
221 host name resolution systems continue to work, such as mDNS,
222 LDAP, etc. This API is based on libasyncns, but it has been
223 cleaned up for inclusion in systemd.
225 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
226 "sd-daemon.h" are no longer found in individual libraries
227 libsystemd-journal.so, libsystemd-login.so,
228 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
229 merged them into a single library, libsystemd.so, which
230 provides all symbols. The reason for this is cyclic
231 dependencies, as these libraries tend to use each other's
232 symbols. So far, we've managed to workaround that by linking
233 a copy of a good part of our code into each of these
234 libraries again and again, which, however, makes certain
235 things hard to do, like sharing static variables. Also, it
236 substantially increases footprint. With this change, there
237 is only one library for the basic APIs systemd
238 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
239 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
240 library as well, however are subject to the --enable-kdbus
241 switch (see below). Note that "sd-dhcp-client.h" is not part
242 of this library (this is because it only consumes, never
243 provides, services of/to other APIs). To make the transition
244 easy from the separate libraries to the unified one, we
245 provide the --enable-compat-libs compile-time switch which
246 will generate stub libraries that are compatible with the
247 old ones but redirect all calls to the new one.
249 * All of the kdbus logic and the new APIs "sd-bus.h",
250 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
251 and "sd-utf8.h" are compile-time optional via the
252 "--enable-kdbus" switch, and they are not compiled in by
253 default. To make use of kdbus, you have to explicitly enable
254 the switch. Note however, that neither the kernel nor the
255 userspace API for all of this is considered stable yet. We
256 want to maintain the freedom to still change the APIs for
257 now. By specifying this build-time switch, you acknowledge
258 that you are aware of the instability of the current
261 * Also, note that while kdbus is pretty much complete,
262 it lacks one thing: proper policy support. This means you
263 can build a fully working system with all features; however,
264 it will be highly insecure. Policy support will be added in
265 one of the next releases, at the same time that we will
266 declare the APIs stable.
268 * When the kernel command-line argument "kdbus" is specified,
269 systemd will automatically load the kdbus.ko kernel module. At
270 this stage of development, it is only useful for testing kdbus
271 and should not be used in production. Note: if "--enable-kdbus"
272 is specified, and the kdbus.ko kernel module is available, and
273 "kdbus" is added to the kernel command line, the entire system
274 runs with kdbus instead of dbus-daemon, with the above mentioned
275 problem of missing the system policy enforcement. Also a future
276 version of kdbus.ko or a newer systemd will not be compatible with
277 each other, and will unlikely be able to boot the machine if only
278 one of them is updated.
280 * systemctl gained a new "import-environment" command which
281 uploads the caller's environment (or parts thereof) into the
282 service manager so that it is inherited by services started
283 by the manager. This is useful to upload variables like
284 $DISPLAY into the user service manager.
286 * A new PrivateDevices= switch has been added to service units
287 which allows running a service with a namespaced /dev
288 directory that does not contain any device nodes for
289 physical devices. More specifically, it only includes devices
290 such as /dev/null, /dev/urandom, and /dev/zero which are API
293 * logind has been extended to support behaviour like VT
294 switching on seats that do not support a VT. This makes
295 multi-session available on seats that are not the first seat
296 (seat0), and on systems where kernel support for VTs has
297 been disabled at compile-time.
299 * If a process holds a delay lock for system sleep or shutdown
300 and fails to release it in time, we will now log its
301 identity. This makes it easier to identify processes that
302 cause slow suspends or power-offs.
304 * When parsing /etc/crypttab, support for a new key-slot=
305 option as supported by Debian is added. It allows indicating
306 which LUKS slot to use on disk, speeding up key loading.
308 * The sd_journald_sendv() API call has been checked and
309 officially declared to be async-signal-safe so that it may
310 be invoked from signal handlers for logging purposes.
312 * Boot-time status output is now enabled automatically after a
313 short timeout if boot does not progress, in order to give
314 the user an indication what she or he is waiting for.
316 * The boot-time output has been improved to show how much time
317 remains until jobs expire.
319 * The KillMode= switch in service units gained a new possible
320 value "mixed". If set, and the unit is shut down, then the
321 initial SIGTERM signal is sent only to the main daemon
322 process, while the following SIGKILL signal is sent to
323 all remaining processes of the service.
325 * When a scope unit is registered, a new property "Controller"
326 may be set. If set to a valid bus name, systemd will send a
327 RequestStop() signal to this name when it would like to shut
328 down the scope. This may be used to hook manager logic into
329 the shutdown logic of scope units. Also, scope units may now
330 be put in a special "abandoned" state, in which case the
331 manager process which created them takes no further
332 responsibilities for it.
334 * When reading unit files, systemd will now verify
335 the access mode of these files, and warn about certain
336 suspicious combinations. This has been added to make it
337 easier to track down packaging bugs where unit files are
338 marked executable or world-writable.
340 * systemd-nspawn gained a new "--setenv=" switch to set
341 container-wide environment variables. The similar option in
342 systemd-activate was renamed from "--environment=" to
343 "--setenv=" for consistency.
345 * systemd-nspawn has been updated to create a new kdbus domain
346 for each container that is invoked, thus allowing each
347 container to have its own set of system and user buses,
348 independent of the host.
350 * systemd-nspawn gained a new --drop-capability= switch to run
351 the container with less capabilities than the default. Both
352 --drop-capability= and --capability= now take the special
353 string "all" for dropping or keeping all capabilities.
355 * systemd-nspawn gained new switches for executing containers
356 with specific SELinux labels set.
358 * systemd-nspawn gained a new --quiet switch to not generate
359 any additional output but the container's own console
362 * systemd-nspawn gained a new --share-system switch to run a
363 container without PID namespacing enabled.
365 * systemd-nspawn gained a new --register= switch to control
366 whether the container is registered with systemd-machined or
367 not. This is useful for containers that do not run full
368 OS images, but only specific apps.
370 * systemd-nspawn gained a new --keep-unit which may be used
371 when invoked as the only program from a service unit, and
372 results in registration of the unit service itself in
373 systemd-machined, instead of a newly opened scope unit.
375 * systemd-nspawn gained a new --network-interface= switch for
376 moving arbitrary interfaces to the container. The new
377 --network-veth switch creates a virtual Ethernet connection
378 between host and container. The new --network-bridge=
379 switch then allows assigning the host side of this virtual
380 Ethernet connection to a bridge device.
382 * systemd-nspawn gained a new --personality= switch for
383 setting the kernel personality for the container. This is
384 useful when running a 32bit container on a 64bit host. A
385 similar option Personality= is now also available in service
388 * logind will now also track a "Desktop" identifier for each
389 session which encodes the desktop environment of it. This is
390 useful for desktop environments that want to identify
391 multiple running sessions of itself easily.
393 * A new SELinuxContext= setting for service units has been
394 added that allows setting a specific SELinux execution
395 context for a service.
397 * Most systemd client tools will now honour $SYSTEMD_LESS for
398 settings of the "less" pager. By default, these tools will
399 override $LESS to allow certain operations to work, such as
400 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
401 influence this logic.
403 * systemd's "seccomp" hook-up has been changed to make use of
404 the libseccomp library instead of using its own
405 implementation. This has benefits for portability among
408 * For usage together with SystemCallFilter=, a new
409 SystemCallErrorNumber= setting has been introduced that
410 allows configuration of a system error number to return on
411 filtered system calls, instead of immediately killing the
412 process. Also, SystemCallArchitectures= has been added to
413 limit access to system calls of a particular architecture
414 (in order to turn off support for unused secondary
415 architectures). There is also a global
416 SystemCallArchitectures= setting in system.conf now to turn
417 off support for non-native system calls system-wide.
419 * systemd requires a kernel with a working name_to_handle_at(),
420 please see the kernel config requirements in the README file.
422 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
423 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
424 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
425 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
426 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
427 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
428 Elia Pinto, Florian Weimer, George McCollister, Goffredo
429 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
430 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
431 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
432 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
433 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
434 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
435 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
436 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
437 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
438 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
439 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
440 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
441 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
442 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
443 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
444 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
445 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
447 -- Berlin, 2014-02-20
451 * logind has gained support for facilitating privileged input
452 and drm device access for unprivileged clients. This work is
453 useful to allow Wayland display servers (and similar
454 programs, such as kmscon) to run under the user's ID and
455 access input and drm devices which are normally
456 protected. When this is used (and the kernel is new enough)
457 logind will "mute" IO on the file descriptors passed to
458 Wayland as long as it is in the background and "unmute" it
459 if it returns into the foreground. This allows secure
460 session switching without allowing background sessions to
461 eavesdrop on input and display data. This also introduces
462 session switching support if VT support is turned off in the
463 kernel, and on seats that are not seat0.
465 * A new kernel command line option luks.options= is understood
466 now which allows specifiying LUKS options for usage for LUKS
467 encrypted partitions specified with luks.uuid=.
469 * tmpfiles.d(5) snippets may now use specifier expansion in
470 path names. More specifically %m, %b, %H, %v, are now
471 replaced by the local machine id, boot id, hostname, and
472 kernel version number.
474 * A new tmpfiles.d(5) command "m" has been introduced which
475 may be used to change the owner/group/access mode of a file
476 or directory if it exists, but do nothing if it doesn't.
478 * This release removes high-level support for the
479 MemorySoftLimit= cgroup setting. The underlying kernel
480 cgroup attribute memory.soft_limit= is currently badly
481 designed and likely to be removed from the kernel API in its
482 current form, hence we shouldn't expose it for now.
484 * The memory.use_hierarchy cgroup attribute is now enabled for
485 all cgroups systemd creates in the memory cgroup
486 hierarchy. This option is likely to be come the built-in
487 default in the kernel anyway, and the non-hierarchial mode
488 never made much sense in the intrinsically hierarchial
491 * A new field _SYSTEMD_SLICE= is logged along with all journal
492 messages containing the slice a message was generated
493 from. This is useful to allow easy per-customer filtering of
494 logs among other things.
496 * systemd-journald will no longer adjust the group of journal
497 files it creates to the "systemd-journal" group. Instead we
498 rely on the journal directory to be owned by the
499 "systemd-journal" group, and its setgid bit set, so that the
500 kernel file system layer will automatically enforce that
501 journal files inherit this group assignment. The reason for
502 this change is that we cannot allow NSS look-ups from
503 journald which would be necessary to resolve
504 "systemd-journal" to a numeric GID, because this might
505 create deadlocks if NSS involves synchronous queries to
506 other daemons (such as nscd, or sssd) which in turn are
507 logging clients of journald and might block on it, which
508 would then dead lock. A tmpfiles.d(5) snippet included in
509 systemd will make sure the setgid bit and group are
510 properly set on the journal directory if it exists on every
511 boot. However, we recommend adjusting it manually after
512 upgrades too (or from RPM scriptlets), so that the change is
513 not delayed until next reboot.
515 * Backlight and random seed files in /var/lib/ have moved into
516 the /var/lib/systemd/ directory, in order to centralize all
517 systemd generated files in one directory.
519 * Boot time performance measurements (as displayed by
520 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
521 performance information if that's available to determine how
522 much time BIOS and boot loader initialization required. With
523 a sufficiently new BIOS you hence no longer need to boot
524 with Gummiboot to get access to such information.
526 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
527 Cristian Rodríguez, Dave Reisner, David Herrmann, David
528 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
529 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
530 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
531 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
532 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
534 -- Berlin, 2013-10-02
538 * The Restart= option for services now understands a new
539 on-watchdog setting, which will restart the service
540 automatically if the service stops sending out watchdog keep
541 alive messages (as configured with WatchdogSec=).
543 * The getty generator (which is responsible for bringing up a
544 getty on configured serial consoles) will no longer only
545 start a getty on the primary kernel console but on all
546 others, too. This makes the order in which console= is
547 specified on the kernel command line less important.
549 * libsystemd-logind gained a new sd_session_get_vt() call to
550 retrieve the VT number of a session.
552 * If the option "tries=0" is set for an entry of /etc/crypttab
553 its passphrase is queried indefinitely instead of any
554 maximum number of tries.
556 * If a service with a configure PID file terminates its PID
557 file will now be removed automatically if it still exists
558 afterwards. This should put an end to stale PID files.
560 * systemd-run will now also take relative binary path names
561 for execution and no longer insists on absolute paths.
563 * InaccessibleDirectories= and ReadOnlyDirectories= now take
564 paths that are optionally prefixed with "-" to indicate that
565 it should not be considered a failure if they don't exist.
567 * journalctl -o (and similar commands) now understands a new
568 output mode "short-precise", it is similar to "short" but
569 shows timestamps with usec accuracy.
571 * The option "discard" (as known from Debian) is now
572 synonymous to "allow-discards" in /etc/crypttab. In fact,
573 "discard" is preferred now (since it is easier to remember
576 * Some licensing clean-ups were made, so that more code is now
577 LGPL-2.1 licensed than before.
579 * A minimal tool to save/restore the display backlight
580 brightness across reboots has been added. It will store the
581 backlight setting as late as possible at shutdown, and
582 restore it as early as possible during reboot.
584 * A logic to automatically discover and enable home and swap
585 partitions on GPT disks has been added. With this in place
586 /etc/fstab becomes optional for many setups as systemd can
587 discover certain partitions located on the root disk
588 automatically. Home partitions are recognized under their
589 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
590 partitions are recognized under their GPT type ID
591 0657fd6da4ab43c484e50933c84b4f4f.
593 * systemd will no longer pass any environment from the kernel
594 or initrd to system services. If you want to set an
595 environment for all services, do so via the kernel command
596 line systemd.setenv= assignment.
598 * The systemd-sysctl tool no longer natively reads the file
599 /etc/sysctl.conf. If desired, the file should be symlinked
600 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
601 legacy support by a symlink rather than built-in code, it
602 also makes the otherwise hidden order of application of the
603 different files visible. (Note that this partly reverts to a
604 pre-198 application order of sysctl knobs!)
606 * The "systemctl set-log-level" and "systemctl dump" commands
607 have been moved to systemd-analyze.
609 * systemd-run learned the new --remain-after-exit switch,
610 which causes the scope unit not to be cleaned up
611 automatically after the process terminated.
613 * tmpfiles learned a new --exclude-prefix= switch to exclude
614 certain paths from operation.
616 * journald will now automatically flush all messages to disk
617 as soon as a message of the log priorities CRIT, ALERT or
620 Contributions from: Andrew Cook, Brandon Philips, Christian
621 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
622 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
623 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
624 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
625 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
626 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
627 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
628 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
629 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
630 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
631 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
632 William Giokas, Zbigniew Jędrzejewski-Szmek
634 -- Berlin, 2013-09-13
638 * The documentation has been updated to cover the various new
639 concepts introduced with 205.
641 * Unit files now understand the new %v specifier which
642 resolves to the kernel version string as returned by "uname
645 * systemctl now supports filtering the unit list output by
646 load state, active state and sub state, using the new
649 * "systemctl status" will now show the results of the
650 condition checks (like ConditionPathExists= and similar) of
651 the last start attempts of the unit. They are also logged to
654 * "journalctl -b" may now be used to look for boot output of a
655 specific boot. Try "journalctl -b -1" for the previous boot,
656 but the syntax is substantially more powerful.
658 * "journalctl --show-cursor" has been added which prints the
659 cursor string the last shown log line. This may then be used
660 with the new "journalctl --after-cursor=" switch to continue
661 browsing logs from that point on.
663 * "journalctl --force" may now be used to force regeneration
666 * Creation of "dead" device nodes has been moved from udev
667 into kmod and tmpfiles. Previously, udev would read the kmod
668 databases to pre-generate dead device nodes based on meta
669 information contained in kernel modules, so that these would
670 be auto-loaded on access rather then at boot. As this
671 doesn't really have much to do with the exposing actual
672 kernel devices to userspace this has always been slightly
673 alien in the udev codebase. Following the new scheme kmod
674 will now generate a runtime snippet for tmpfiles from the
675 module meta information and it now is tmpfiles' job to the
676 create the nodes. This also allows overriding access and
677 other parameters for the nodes using the usual tmpfiles
678 facilities. As side effect this allows us to remove the
679 CAP_SYS_MKNOD capability bit from udevd entirely.
681 * logind's device ACLs may now be applied to these "dead"
682 devices nodes too, thus finally allowing managed access to
683 devices such as /dev/snd/sequencer whithout loading the
684 backing module right-away.
686 * A new RPM macro has been added that may be used to apply
687 tmpfiles configuration during package installation.
689 * systemd-detect-virt and ConditionVirtualization= now can
690 detect User-Mode-Linux machines (UML).
692 * journald will now implicitly log the effective capabilities
693 set of processes in the message metadata.
695 * systemd-cryptsetup has gained support for TrueCrypt volumes.
697 * The initrd interface has been simplified (more specifically,
698 support for passing performance data via environment
699 variables and fsck results via files in /run has been
700 removed). These features were non-essential, and are
701 nowadays available in a much nicer way by having systemd in
702 the initrd serialize its state and have the hosts systemd
703 deserialize it again.
705 * The udev "keymap" data files and tools to apply keyboard
706 specific mappings of scan to key codes, and force-release
707 scan code lists have been entirely replaced by a udev
708 "keyboard" builtin and a hwdb data file.
710 * systemd will now honour the kernel's "quiet" command line
711 argument also during late shutdown, resulting in a
712 completely silent shutdown when used.
714 * There's now an option to control the SO_REUSEPORT socket
715 option in .socket units.
717 * Instance units will now automatically get a per-template
718 subslice of system.slice unless something else is explicitly
719 configured. For example, instances of sshd@.service will now
720 implicitly be placed in system-sshd.slice rather than
721 system.slice as before.
723 * Test coverage support may now be enabled at build time.
725 Contributions from: Dave Reisner, Frederic Crozat, Harald
726 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
727 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
728 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
729 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
730 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
731 Giokas, Zbigniew Jędrzejewski-Szmek
733 -- Berlin, 2013-07-23
737 * Two new unit types have been introduced:
739 Scope units are very similar to service units, however, are
740 created out of pre-existing processes -- instead of PID 1
741 forking off the processes. By using scope units it is
742 possible for system services and applications to group their
743 own child processes (worker processes) in a powerful way
744 which then maybe used to organize them, or kill them
745 together, or apply resource limits on them.
747 Slice units may be used to partition system resources in an
748 hierarchial fashion and then assign other units to them. By
749 default there are now three slices: system.slice (for all
750 system services), user.slice (for all user sessions),
751 machine.slice (for VMs and containers).
753 Slices and scopes have been introduced primarily in
754 context of the work to move cgroup handling to a
755 single-writer scheme, where only PID 1
756 creates/removes/manages cgroups.
758 * There's a new concept of "transient" units. In contrast to
759 normal units these units are created via an API at runtime,
760 not from configuration from disk. More specifically this
761 means it is now possible to run arbitrary programs as
762 independent services, with all execution parameters passed
763 in via bus APIs rather than read from disk. Transient units
764 make systemd substantially more dynamic then it ever was,
765 and useful as a general batch manager.
767 * logind has been updated to make use of scope and slice units
768 for managing user sessions. As a user logs in he will get
769 his own private slice unit, to which all sessions are added
770 as scope units. We also added support for automatically
771 adding an instance of user@.service for the user into the
772 slice. Effectively logind will no longer create cgroup
773 hierarchies on its own now, it will defer entirely to PID 1
774 for this by means of scope, service and slice units. Since
775 user sessions this way become entities managed by PID 1
776 the output of "systemctl" is now a lot more comprehensive.
778 * A new mini-daemon "systemd-machined" has been added which
779 may be used by virtualization managers to register local
780 VMs/containers. nspawn has been updated accordingly, and
781 libvirt will be updated shortly. machined will collect a bit
782 of meta information about the VMs/containers, and assign
783 them their own scope unit (see above). The collected
784 meta-data is then made available via the "machinectl" tool,
785 and exposed in "ps" and similar tools. machined/machinectl
786 is compile-time optional.
788 * As discussed earlier, the low-level cgroup configuration
789 options ControlGroup=, ControlGroupModify=,
790 ControlGroupPersistent=, ControlGroupAttribute= have been
791 removed. Please use high-level attribute settings instead as
794 * A new bus call SetUnitProperties() has been added to alter
795 various runtime parameters of a unit. This is primarily
796 useful to alter cgroup parameters dynamically in a nice way,
797 but will be extended later on to make more properties
798 modifiable at runtime. systemctl gained a new set-properties
799 command that wraps this call.
801 * A new tool "systemd-run" has been added which can be used to
802 run arbitrary command lines as transient services or scopes,
803 while configuring a number of settings via the command
804 line. This tool is currently very basic, however already
805 very useful. We plan to extend this tool to even allow
806 queuing of execution jobs with time triggers from the
807 command line, similar in fashion to "at".
809 * nspawn will now inform the user explicitly that kernels with
810 audit enabled break containers, and suggest the user to turn
813 * Support for detecting the IMA and AppArmor security
814 frameworks with ConditionSecurity= has been added.
816 * journalctl gained a new "-k" switch for showing only kernel
817 messages, mimicking dmesg output; in addition to "--user"
818 and "--system" switches for showing only user's own logs
821 * systemd-delta can now show information about drop-in
822 snippets extending unit files.
824 * libsystemd-bus has been substantially updated but is still
825 not available as public API.
827 * systemd will now look for the "debug" argument on the kernel
828 command line and enable debug logging, similar to
829 "systemd.log_level=debug" already did before.
831 * "systemctl set-default", "systemctl get-default" has been
832 added to configure the default.target symlink, which
833 controls what to boot into by default.
835 * "systemctl set-log-level" has been added as a convenient
836 way to raise and lower systemd logging threshold.
838 * "systemd-analyze plot" will now show the time the various
839 generators needed for execution, as well as information
840 about the unit file loading.
842 * libsystemd-journal gained a new sd_journal_open_files() call
843 for opening specific journal files. journactl also gained a
844 new switch to expose this new functionality. Previously we
845 only supported opening all files from a directory, or all
846 files from the system, as opening individual files only is
847 racy due to journal file rotation.
849 * systemd gained the new DefaultEnvironment= setting in
850 /etc/systemd/system.conf to set environment variables for
853 * If a privileged process logs a journal message with the
854 OBJECT_PID= field set, then journald will automatically
855 augment this with additional OBJECT_UID=, OBJECT_GID=,
856 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
857 system services want to log events about specific client
858 processes. journactl/systemctl has been updated to make use
859 of this information if all log messages regarding a specific
862 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
863 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
864 Reisner, David Coppa, David King, David Strauss, Eelco
865 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
866 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
867 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
868 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
869 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
870 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
871 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
872 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
873 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
878 * The Python bindings gained some minimal support for the APIs
879 exposed by libsystemd-logind.
881 * ConditionSecurity= gained support for detecting SMACK. Since
882 this condition already supports SELinux and AppArmor we only
883 miss IMA for this. Patches welcome!
885 Contributions from: Karol Lewandowski, Lennart Poettering,
886 Zbigniew Jędrzejewski-Szmek
890 * systemd-nspawn will now create /etc/resolv.conf if
891 necessary, before bind-mounting the host's file onto it.
893 * systemd-nspawn will now store meta information about a
894 container on the container's cgroup as extended attribute
895 fields, including the root directory.
897 * The cgroup hierarchy has been reworked in many ways. All
898 objects any of the components systemd creates in the cgroup
899 tree are now suffixed. More specifically, user sessions are
900 now placed in cgroups suffixed with ".session", users in
901 cgroups suffixed with ".user", and nspawn containers in
902 cgroups suffixed with ".nspawn". Furthermore, all cgroup
903 names are now escaped in a simple scheme to avoid collision
904 of userspace object names with kernel filenames. This work
905 is preparation for making these objects relocatable in the
906 cgroup tree, in order to allow easy resource partitioning of
907 these objects without causing naming conflicts.
909 * systemctl list-dependencies gained the new switches
910 --plain, --reverse, --after and --before.
912 * systemd-inhibit now shows the process name of processes that
913 have taken an inhibitor lock.
915 * nss-myhostname will now also resolve "localhost"
916 implicitly. This makes /etc/hosts an optional file and
917 nicely handles that on IPv6 ::1 maps to both "localhost" and
920 * libsystemd-logind.so gained a new call
921 sd_get_machine_names() to enumerate running containers and
922 VMs (currently only supported by very new libvirt and
923 nspawn). sd_login_monitor can now be used to watch
924 VMs/containers coming and going.
926 * .include is not allowed recursively anymore, and only in
927 unit files. Usually it is better to use drop-in snippets in
928 .d/*.conf anyway, as introduced with systemd 198.
930 * systemd-analyze gained a new "critical-chain" command that
931 determines the slowest chain of units run during system
932 boot-up. It is very useful for tracking down where
933 optimizing boot time is the most beneficial.
935 * systemd will no longer allow manipulating service paths in
936 the name=systemd:/system cgroup tree using ControlGroup= in
937 units. (But is still fine with it in all other dirs.)
939 * There's a new systemd-nspawn@.service service file that may
940 be used to easily run nspawn containers as system
941 services. With the container's root directory in
942 /var/lib/container/foobar it is now sufficient to run
943 "systemctl start systemd-nspawn@foobar.service" to boot it.
945 * systemd-cgls gained a new parameter "--machine" to list only
946 the processes within a certain container.
948 * ConditionSecurity= now can check for "apparmor". We still
949 are lacking checks for SMACK and IMA for this condition
950 check though. Patches welcome!
952 * A new configuration file /etc/systemd/sleep.conf has been
953 added that may be used to configure which kernel operation
954 systemd is supposed to execute when "suspend", "hibernate"
955 or "hybrid-sleep" is requested. This makes the new kernel
956 "freeze" state accessible to the user.
958 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
959 the passed argument if applicable.
961 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
962 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
963 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
964 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
965 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
966 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
967 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
972 * The output of 'systemctl list-jobs' got some polishing. The
973 '--type=' argument may now be passed more than once. A new
974 command 'systemctl list-sockets' has been added which shows
975 a list of kernel sockets systemd is listening on with the
976 socket units they belong to, plus the units these socket
979 * The experimental libsystemd-bus library got substantial
980 updates to work in conjunction with the (also experimental)
981 kdbus kernel project. It works well enough to exchange
982 messages with some sophistication. Note that kdbus is not
983 ready yet, and the library is mostly an elaborate test case
984 for now, and not installable.
986 * systemd gained a new unit 'systemd-static-nodes.service'
987 that generates static device nodes earlier during boot, and
988 can run in conjunction with udev.
990 * libsystemd-login gained a new call sd_pid_get_user_unit()
991 to retrieve the user systemd unit a process is running
992 in. This is useful for systems where systemd is used as
995 * systemd-nspawn now places all containers in the new /machine
996 top-level cgroup directory in the name=systemd
997 hierarchy. libvirt will soon do the same, so that we get a
998 uniform separation of /system, /user and /machine for system
999 services, user processes and containers/virtual
1000 machines. This new cgroup hierarchy is also useful to stick
1001 stable names to specific container instances, which can be
1002 recognized later this way (this name may be controlled
1003 via systemd-nspawn's new -M switch). libsystemd-login also
1004 gained a new call sd_pid_get_machine_name() to retrieve the
1005 name of the container/VM a specific process belongs to.
1007 * bootchart can now store its data in the journal.
1009 * libsystemd-journal gained a new call
1010 sd_journal_add_conjunction() for AND expressions to the
1011 matching logic. This can be used to express more complex
1012 logical expressions.
1014 * journactl can now take multiple --unit= and --user-unit=
1017 * The cryptsetup logic now understands the "luks.key=" kernel
1018 command line switch for specifying a file to read the
1019 decryption key from. Also, if a configured key file is not
1020 found the tool will now automatically fall back to prompting
1023 * Python systemd.journal module was updated to wrap recently
1024 added functions from libsystemd-journal. The interface was
1025 changed to bring the low level interface in s.j._Reader
1026 closer to the C API, and the high level interface in
1027 s.j.Reader was updated to wrap and convert all data about
1030 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
1031 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
1032 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
1033 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
1034 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
1035 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1039 * journalctl --update-catalog now understands a new --root=
1040 option to operate on catalogs found in a different root
1043 * During shutdown after systemd has terminated all running
1044 services a final killing loop kills all remaining left-over
1045 processes. We will now print the name of these processes
1046 when we send SIGKILL to them, since this usually indicates a
1049 * If /etc/crypttab refers to password files stored on
1050 configured mount points automatic dependencies will now be
1051 generated to ensure the specific mount is established first
1052 before the key file is attempted to be read.
1054 * 'systemctl status' will now show information about the
1055 network sockets a socket unit is listening on.
1057 * 'systemctl status' will also shown information about any
1058 drop-in configuration file for units. (Drop-In configuration
1059 files in this context are files such as
1060 /etc/systemd/systemd/foobar.service.d/*.conf)
1062 * systemd-cgtop now optionally shows summed up CPU times of
1063 cgroups. Press '%' while running cgtop to switch between
1064 percentage and absolute mode. This is useful to determine
1065 which cgroups use up the most CPU time over the entire
1066 runtime of the system. systemd-cgtop has also been updated
1067 to be 'pipeable' for processing with further shell tools.
1069 * 'hostnamectl set-hostname' will now allow setting of FQDN
1072 * The formatting and parsing of time span values has been
1073 changed. The parser now understands fractional expressions
1074 such as "5.5h". The formatter will now output fractional
1075 expressions for all time spans under 1min, i.e. "5.123456s"
1076 rather than "5s 123ms 456us". For time spans under 1s
1077 millisecond values are shown, for those under 1ms
1078 microsecond values are shown. This should greatly improve
1079 all time-related output of systemd.
1081 * libsystemd-login and libsystemd-journal gained new
1082 functions for querying the poll() events mask and poll()
1083 timeout value for integration into arbitrary event
1086 * localectl gained the ability to list available X11 keymaps
1087 (models, layouts, variants, options).
1089 * 'systemd-analyze dot' gained the ability to filter for
1090 specific units via shell-style globs, to create smaller,
1091 more useful graphs. I.e. it's now possible to create simple
1092 graphs of all the dependencies between only target units, or
1093 of all units that Avahi has dependencies with.
1095 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
1096 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
1097 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
1098 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
1099 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
1100 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
1101 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
1105 * The boot-time readahead implementation for rotating media
1106 will now read the read-ahead data in multiple passes which
1107 consist of all read requests made in equidistant time
1108 intervals. This means instead of strictly reading read-ahead
1109 data in its physical order on disk we now try to find a
1110 middle ground between physical and access time order.
1112 * /etc/os-release files gained a new BUILD_ID= field for usage
1113 on operating systems that provide continuous builds of OS
1116 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
1117 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
1118 William Douglas, Zbigniew Jędrzejewski-Szmek
1122 * systemd-python gained an API exposing libsystemd-daemon.
1124 * The SMACK setup logic gained support for uploading CIPSO
1127 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
1128 ReadOnlyDirectories= and InaccessibleDirectories= has
1129 changed. The private /tmp and /var/tmp directories are now
1130 shared by all processes of a service (which means
1131 ExecStartPre= may now leave data in /tmp that ExecStart= of
1132 the same service can still access). When a service is
1133 stopped its temporary directories are immediately deleted
1134 (normal clean-up with tmpfiles is still done in addition to
1137 * By default, systemd will now set a couple of sysctl
1138 variables in the kernel: the safe sysrq options are turned
1139 on, IP route verification is turned on, and source routing
1140 disabled. The recently added hardlink and softlink
1141 protection of the kernel is turned on. These settings should
1142 be reasonably safe, and good defaults for all new systems.
1144 * The predictable network naming logic may now be turned off
1145 with a new kernel command line switch: net.ifnames=0.
1147 * A new libsystemd-bus module has been added that implements a
1148 pretty complete D-Bus client library. For details see:
1150 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
1152 * journald will now explicitly flush the journal files to disk
1153 at the latest 5min after each write. The file will then also
1154 be marked offline until the next write. This should increase
1155 reliability in case of a crash. The synchronization delay
1156 can be configured via SyncIntervalSec= in journald.conf.
1158 * There's a new remote-fs-setup.target unit that can be used
1159 to pull in specific services when at least one remote file
1160 system is to be mounted.
1162 * There are new targets timers.target and paths.target as
1163 canonical targets to pull user timer and path units in
1164 from. This complements sockets.target with a similar
1165 purpose for socket units.
1167 * libudev gained a new call udev_device_set_attribute_value()
1168 to set sysfs attributes of a device.
1170 * The udev daemon now sets the default number of worker
1171 processes executed in parallel based on the number of available
1172 CPUs instead of the amount of available RAM. This is supposed
1173 to provide a more reliable default and limit a too aggressive
1174 paralellism for setups with 1000s of devices connected.
1176 Contributions from: Auke Kok, Colin Walters, Cristian
1177 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
1178 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
1179 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
1180 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
1181 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
1182 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
1183 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
1184 Zbigniew Jędrzejewski-Szmek
1188 * Configuration of unit files may now be extended via drop-in
1189 files without having to edit/override the unit files
1190 themselves. More specifically, if the administrator wants to
1191 change one value for a service file foobar.service he can
1192 now do so by dropping in a configuration snippet into
1193 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
1194 will load all these snippets and apply them on top of the
1195 main unit configuration file, possibly extending or
1196 overriding its settings. Using these drop-in snippets is
1197 generally nicer than the two earlier options for changing
1198 unit files locally: copying the files from
1199 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
1200 them there; or creating a new file in /etc/systemd/system/
1201 that incorporates the original one via ".include". Drop-in
1202 snippets into these .d/ directories can be placed in any
1203 directory systemd looks for units in, and the usual
1204 overriding semantics between /usr/lib, /etc and /run apply
1207 * Most unit file settings which take lists of items can now be
1208 reset by assigning the empty string to them. For example,
1209 normally, settings such as Environment=FOO=BAR append a new
1210 environment variable assignment to the environment block,
1211 each time they are used. By assigning Environment= the empty
1212 string the environment block can be reset to empty. This is
1213 particularly useful with the .d/*.conf drop-in snippets
1214 mentioned above, since this adds the ability to reset list
1215 settings from vendor unit files via these drop-ins.
1217 * systemctl gained a new "list-dependencies" command for
1218 listing the dependencies of a unit recursively.
1220 * Inhibitors are now honored and listed by "systemctl
1221 suspend", "systemctl poweroff" (and similar) too, not only
1222 GNOME. These commands will also list active sessions by
1225 * Resource limits (as exposed by the various control group
1226 controllers) can now be controlled dynamically at runtime
1227 for all units. More specifically, you can now use a command
1228 like "systemctl set-cgroup-attr foobar.service cpu.shares
1229 2000" to alter the CPU shares a specific service gets. These
1230 settings are stored persistently on disk, and thus allow the
1231 administrator to easily adjust the resource usage of
1232 services with a few simple commands. This dynamic resource
1233 management logic is also available to other programs via the
1234 bus. Almost any kernel cgroup attribute and controller is
1237 * systemd-vconsole-setup will now copy all font settings to
1238 all allocated VTs, where it previously applied them only to
1241 * libsystemd-login gained the new sd_session_get_tty() API
1244 * This release drops support for a few legacy or
1245 distribution-specific LSB facility names when parsing init
1246 scripts: $x-display-manager, $mail-transfer-agent,
1247 $mail-transport-agent, $mail-transfer-agent, $smtp,
1248 $null. Also, the mail-transfer-agent.target unit backing
1249 this has been removed. Distributions which want to retain
1250 compatibility with this should carry the burden for
1251 supporting this themselves and patch support for these back
1252 in, if they really need to. Also, the facilities $syslog and
1253 $local_fs are now ignored, since systemd does not support
1254 early-boot LSB init scripts anymore, and these facilities
1255 are implied anyway for normal services. syslog.target has
1258 * There are new bus calls on PID1's Manager object for
1259 cancelling jobs, and removing snapshot units. Previously,
1260 both calls were only available on the Job and Snapshot
1263 * systemd-journal-gatewayd gained SSL support.
1265 * The various "environment" files, such as /etc/locale.conf
1266 now support continuation lines with a backslash ("\") as
1267 last character in the line, similar in style (but different)
1268 to how this is supported in shells.
1270 * For normal user processes the _SYSTEMD_USER_UNIT= field is
1271 now implicitly appended to every log entry logged. systemctl
1272 has been updated to filter by this field when operating on a
1273 user systemd instance.
1275 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
1276 CAP_AUDIT_CONTROL capabilities to the capabilities set for
1277 the container. This makes it easier to boot unmodified
1278 Fedora systems in a container, which however still requires
1279 audit=0 to be passed on the kernel command line. Auditing in
1280 kernel and userspace is unfortunately still too broken in
1281 context of containers, hence we recommend compiling it out
1282 of the kernel or using audit=0. Hopefully this will be fixed
1283 one day for good in the kernel.
1285 * nspawn gained the new --bind= and --bind-ro= parameters to
1286 bind mount specific directories from the host into the
1289 * nspawn will now mount its own devpts file system instance
1290 into the container, in order not to leak pty devices from
1291 the host into the container.
1293 * systemd will now read the firmware boot time performance
1294 information from the EFI variables, if the used boot loader
1295 supports this, and takes it into account for boot performance
1296 analysis via "systemd-analyze". This is currently supported
1297 only in conjunction with Gummiboot, but could be supported
1298 by other boot loaders too. For details see:
1300 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
1302 * A new generator has been added that automatically mounts the
1303 EFI System Partition (ESP) to /boot, if that directory
1304 exists, is empty, and no other file system has been
1305 configured to be mounted there.
1307 * logind will now send out PrepareForSleep(false) out
1308 unconditionally, after coming back from suspend. This may be
1309 used by applications as asynchronous notification for
1310 system resume events.
1312 * "systemctl unlock-sessions" has been added, that allows
1313 unlocking the screens of all user sessions at once, similar
1314 how "systemctl lock-sessions" already locked all users
1315 sessions. This is backed by a new D-Bus call UnlockSessions().
1317 * "loginctl seat-status" will now show the master device of a
1318 seat. (i.e. the device of a seat that needs to be around for
1319 the seat to be considered available, usually the graphics
1322 * tmpfiles gained a new "X" line type, that allows
1323 configuration of files and directories (with wildcards) that
1324 shall be excluded from automatic cleanup ("aging").
1326 * udev default rules set the device node permissions now only
1327 at "add" events, and do not change them any longer with a
1328 later "change" event.
1330 * The log messages for lid events and power/sleep keypresses
1331 now carry a message ID.
1333 * We now have a substantially larger unit test suite, but this
1334 continues to be work in progress.
1336 * udevadm hwdb gained a new --root= parameter to change the
1337 root directory to operate relative to.
1339 * logind will now issue a background sync() request to the kernel
1340 early at shutdown, so that dirty buffers are flushed to disk early
1341 instead of at the last moment, in order to optimize shutdown
1344 * A new bootctl tool has been added that is an interface for
1345 certain boot loader operations. This is currently a preview
1346 and is likely to be extended into a small mechanism daemon
1347 like timedated, localed, hostnamed, and can be used by
1348 graphical UIs to enumerate available boot options, and
1349 request boot into firmware operations.
1351 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
1352 the rest of the package. It also has been updated to work
1353 correctly in initrds.
1355 * Policykit previously has been runtime optional, and is now
1356 also compile time optional via a configure switch.
1358 * systemd-analyze has been reimplemented in C. Also "systemctl
1359 dot" has moved into systemd-analyze.
1361 * "systemctl status" with no further parameters will now print
1362 the status of all active or failed units.
1364 * Operations such as "systemctl start" can now be executed
1365 with a new mode "--irreversible" which may be used to queue
1366 operations that cannot accidentally be reversed by a later
1367 job queuing. This is by default used to make shutdown
1368 requests more robust.
1370 * The Python API of systemd now gained a new module for
1371 reading journal files.
1373 * A new tool kernel-install has been added that can install
1374 kernel images according to the Boot Loader Specification:
1376 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
1378 * Boot time console output has been improved to provide
1379 animated boot time output for hanging jobs.
1381 * A new tool systemd-activate has been added which can be used
1382 to test socket activation with, directly from the command
1383 line. This should make it much easier to test and debug
1384 socket activation in daemons.
1386 * journalctl gained a new "--reverse" (or -r) option to show
1387 journal output in reverse order (i.e. newest line first).
1389 * journalctl gained a new "--pager-end" (or -e) option to jump
1390 to immediately jump to the end of the journal in the
1391 pager. This is only supported in conjunction with "less".
1393 * journalctl gained a new "--user-unit=" option, that works
1394 similar to "--unit=" but filters for user units rather than
1397 * A number of unit files to ease adoption of systemd in
1398 initrds has been added. This moves some minimal logic from
1399 the various initrd implementations into systemd proper.
1401 * The journal files are now owned by a new group
1402 "systemd-journal", which exists specifically to allow access
1403 to the journal, and nothing else. Previously, we used the
1404 "adm" group for that, which however possibly covers more
1405 than just journal/log file access. This new group is now
1406 already used by systemd-journal-gatewayd to ensure this
1407 daemon gets access to the journal files and as little else
1408 as possible. Note that "make install" will also set FS ACLs
1409 up for /var/log/journal to give "adm" and "wheel" read
1410 access to it, in addition to "systemd-journal" which owns
1411 the journal files. We recommend that packaging scripts also
1412 add read access to "adm" + "wheel" to /var/log/journal, and
1413 all existing/future journal files. To normal users and
1414 administrators little changes, however packagers need to
1415 ensure to create the "systemd-journal" system group at
1416 package installation time.
1418 * The systemd-journal-gatewayd now runs as unprivileged user
1419 systemd-journal-gateway:systemd-journal-gateway. Packaging
1420 scripts need to create these system user/group at
1423 * timedated now exposes a new boolean property CanNTP that
1424 indicates whether a local NTP service is available or not.
1426 * systemd-detect-virt will now also detect xen PVs
1428 * The pstore file system is now mounted by default, if it is
1431 * In addition to the SELinux and IMA policies we will now also
1432 load SMACK policies at early boot.
1434 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
1435 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
1436 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
1437 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
1438 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
1439 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
1440 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
1441 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
1442 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
1443 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
1444 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
1445 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
1446 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
1447 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
1451 * Timer units now support calendar time events in addition to
1452 monotonic time events. That means you can now trigger a unit
1453 based on a calendar time specification such as "Thu,Fri
1454 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
1455 or fifth day of any month of the year 2013, given that it is
1456 a thursday or friday. This brings timer event support
1457 considerably closer to cron's capabilities. For details on
1458 the supported calendar time specification language see
1461 * udev now supports a number of different naming policies for
1462 network interfaces for predictable names, and a combination
1463 of these policies is now the default. Please see this wiki
1464 document for details:
1466 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
1468 * Auke Kok's bootchart implementation has been added to the
1469 systemd tree. It's an optional component that can graph the
1470 boot in quite some detail. It's one of the best bootchart
1471 implementations around and minimal in its code and
1474 * nss-myhostname has been integrated into the systemd source
1475 tree. nss-myhostname guarantees that the local hostname
1476 always stays resolvable via NSS. It has been a weak
1477 requirement of systemd-hostnamed since a long time, and
1478 since its code is actually trivial we decided to just
1479 include it in systemd's source tree. It can be turned off
1480 with a configure switch.
1482 * The read-ahead logic is now capable of properly detecting
1483 whether a btrfs file system is on SSD or rotating media, in
1484 order to optimize the read-ahead scheme. Previously, it was
1485 only capable of detecting this on traditional file systems
1488 * In udev, additional device properties are now read from the
1489 IAB in addition to the OUI database. Also, Bluetooth company
1490 identities are attached to the devices as well.
1492 * In service files %U may be used as specifier that is
1493 replaced by the configured user name of the service.
1495 * nspawn may now be invoked without a controlling TTY. This
1496 makes it suitable for invocation as its own service. This
1497 may be used to set up a simple containerized server system
1498 using only core OS tools.
1500 * systemd and nspawn can now accept socket file descriptors
1501 when they are started for socket activation. This enables
1502 implementation of socket activated nspawn
1503 containers. i.e. think about autospawning an entire OS image
1504 when the first SSH or HTTP connection is received. We expect
1505 that similar functionality will also be added to libvirt-lxc
1508 * journalctl will now suppress ANSI color codes when
1509 presenting log data.
1511 * systemctl will no longer show control group information for
1512 a unit if a the control group is empty anyway.
1514 * logind can now automatically suspend/hibernate/shutdown the
1517 * /etc/machine-info and hostnamed now also expose the chassis
1518 type of the system. This can be used to determine whether
1519 the local system is a laptop, desktop, handset or
1520 tablet. This information may either be configured by the
1521 user/vendor or is automatically determined from ACPI and DMI
1522 information if possible.
1524 * A number of PolicyKit actions are now bound together with
1525 "imply" rules. This should simplify creating UIs because
1526 many actions will now authenticate similar ones as well.
1528 * Unit files learnt a new condition ConditionACPower= which
1529 may be used to conditionalize a unit depending on whether an
1530 AC power source is connected or not, of whether the system
1531 is running on battery power.
1533 * systemctl gained a new "is-failed" verb that may be used in
1534 shell scripts and suchlike to check whether a specific unit
1535 is in the "failed" state.
1537 * The EnvironmentFile= setting in unit files now supports file
1538 globbing, and can hence be used to easily read a number of
1539 environment files at once.
1541 * systemd will no longer detect and recognize specific
1542 distributions. All distribution-specific #ifdeffery has been
1543 removed, systemd is now fully generic and
1544 distribution-agnostic. Effectively, not too much is lost as
1545 a lot of the code is still accessible via explicit configure
1546 switches. However, support for some distribution specific
1547 legacy configuration file formats has been dropped. We
1548 recommend distributions to simply adopt the configuration
1549 files everybody else uses now and convert the old
1550 configuration from packaging scripts. Most distributions
1551 already did that. If that's not possible or desirable,
1552 distributions are welcome to forward port the specific
1553 pieces of code locally from the git history.
1555 * When logging a message about a unit systemd will now always
1556 log the unit name in the message meta data.
1558 * localectl will now also discover system locale data that is
1559 not stored in locale archives, but directly unpacked.
1561 * logind will no longer unconditionally use framebuffer
1562 devices as seat masters, i.e. as devices that are required
1563 to be existing before a seat is considered preset. Instead,
1564 it will now look for all devices that are tagged as
1565 "seat-master" in udev. By default framebuffer devices will
1566 be marked as such, but depending on local systems other
1567 devices might be marked as well. This may be used to
1568 integrate graphics cards using closed source drivers (such
1569 as NVidia ones) more nicely into logind. Note however, that
1570 we recommend using the open source NVidia drivers instead,
1571 and no udev rules for the closed-source drivers will be
1572 shipped from us upstream.
1574 Contributions from: Adam Williamson, Alessandro Crismani, Auke
1575 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
1576 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
1577 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
1578 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
1579 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
1580 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
1581 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
1582 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
1583 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
1584 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
1585 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
1590 * udev gained support for loading additional device properties
1591 from an indexed database that is keyed by vendor/product IDs
1592 and similar device identifiers. For the beginning this
1593 "hwdb" is populated with data from the well-known PCI and
1594 USB database, but also includes PNP, ACPI and OID data. In
1595 the longer run this indexed database shall grow into
1596 becoming the one central database for non-essential
1597 userspace device metadata. Previously, data from the PCI/USB
1598 database was only attached to select devices, since the
1599 lookup was a relatively expensive operation due to O(n) time
1600 complexity (with n being the number of entries in the
1601 database). Since this is now O(1), we decided to add in this
1602 data for all devices where this is available, by
1603 default. Note that the indexed database needs to be rebuilt
1604 when new data files are installed. To achieve this you need
1605 to update your packaging scripts to invoke "udevadm hwdb
1606 --update" after installation of hwdb data files. For
1607 RPM-based distributions we introduced the new
1608 %udev_hwdb_update macro for this purpose.
1610 * The Journal gained support for the "Message Catalog", an
1611 indexed database to link up additional information with
1612 journal entries. For further details please check:
1614 http://www.freedesktop.org/wiki/Software/systemd/catalog
1616 The indexed message catalog database also needs to be
1617 rebuilt after installation of message catalog files. Use
1618 "journalctl --update-catalog" for this. For RPM-based
1619 distributions we introduced the %journal_catalog_update
1620 macro for this purpose.
1622 * The Python Journal bindings gained support for the standard
1623 Python logging framework.
1625 * The Journal API gained new functions for checking whether
1626 the underlying file system of a journal file is capable of
1627 properly reporting file change notifications, or whether
1628 applications that want to reflect journal changes "live"
1629 need to recheck journal files continuously in appropriate
1632 * It is now possible to set the "age" field for tmpfiles
1633 entries to 0, indicating that files matching this entry
1634 shall always be removed when the directories are cleaned up.
1636 * coredumpctl gained a new "gdb" verb which invokes gdb
1637 right-away on the selected coredump.
1639 * There's now support for "hybrid sleep" on kernels that
1640 support this, in addition to "suspend" and "hibernate". Use
1641 "systemctl hybrid-sleep" to make use of this.
1643 * logind's HandleSuspendKey= setting (and related settings)
1644 now gained support for a new "lock" setting to simply
1645 request the screen lock on all local sessions, instead of
1646 actually executing a suspend or hibernation.
1648 * systemd will now mount the EFI variables file system by
1651 * Socket units now gained support for configuration of the
1652 SMACK security label.
1654 * timedatectl will now output the time of the last and next
1655 daylight saving change.
1657 * We dropped support for various legacy and distro-specific
1658 concepts, such as insserv, early-boot SysV services
1659 (i.e. those for non-standard runlevels such as 'b' or 'S')
1660 or ArchLinux /etc/rc.conf support. We recommend the
1661 distributions who still need support this to either continue
1662 to maintain the necessary patches downstream, or find a
1663 different solution. (Talk to us if you have questions!)
1665 * Various systemd components will now bypass PolicyKit checks
1666 for root and otherwise handle properly if PolicyKit is not
1667 found to be around. This should fix most issues for
1668 PolicyKit-less systems. Quite frankly this should have been
1669 this way since day one. It is absolutely our intention to
1670 make systemd work fine on PolicyKit-less systems, and we
1671 consider it a bug if something doesn't work as it should if
1672 PolicyKit is not around.
1674 * For embedded systems it is now possible to build udev and
1675 systemd without blkid and/or kmod support.
1677 * "systemctl switch-root" is now capable of switching root
1678 more than once. I.e. in addition to transitions from the
1679 initrd to the host OS it is now possible to transition to
1680 further OS images from the host. This is useful to implement
1681 offline updating tools.
1683 * Various other additions have been made to the RPM macros
1684 shipped with systemd. Use %udev_rules_update() after
1685 installing new udev rules files. %_udevhwdbdir,
1686 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
1687 %_sysctldir are now available which resolve to the right
1688 directories for packages to place various data files in.
1690 * journalctl gained the new --full switch (in addition to
1691 --all, to disable ellipsation for long messages.
1693 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
1694 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
1695 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
1696 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
1697 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
1698 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
1699 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
1700 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
1701 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
1705 * journalctl gained new --since= and --until= switches to
1706 filter by time. It also now supports nice filtering for
1707 units via --unit=/-u.
1709 * Type=oneshot services may use ExecReload= and do the
1712 * The journal daemon now supports time-based rotation and
1713 vacuuming, in addition to the usual disk-space based
1716 * The journal will now index the available field values for
1717 each field name. This enables clients to show pretty drop
1718 downs of available match values when filtering. The bash
1719 completion of journalctl has been updated
1720 accordingly. journalctl gained a new switch -F to list all
1721 values a certain field takes in the journal database.
1723 * More service events are now written as structured messages
1724 to the journal, and made recognizable via message IDs.
1726 * The timedated, localed and hostnamed mini-services which
1727 previously only provided support for changing time, locale
1728 and hostname settings from graphical DEs such as GNOME now
1729 also have a minimal (but very useful) text-based client
1730 utility each. This is probably the nicest way to changing
1731 these settings from the command line now, especially since
1732 it lists available options and is fully integrated with bash
1735 * There's now a new tool "systemd-coredumpctl" to list and
1736 extract coredumps from the journal.
1738 * We now install a README each in /var/log/ and
1739 /etc/rc.d/init.d explaining where the system logs and init
1740 scripts went. This hopefully should help folks who go to
1741 that dirs and look into the otherwise now empty void and
1742 scratch their heads.
1744 * When user-services are invoked (by systemd --user) the
1745 $MANAGERPID env var is set to the PID of systemd.
1747 * SIGRTMIN+24 when sent to a --user instance will now result
1748 in immediate termination of systemd.
1750 * gatewayd received numerous feature additions such as a
1751 "follow" mode, for live syncing and filtering.
1753 * browse.html now allows filtering and showing detailed
1754 information on specific entries. Keyboard navigation and
1755 mouse screen support has been added.
1757 * gatewayd/journalctl now supports HTML5/JSON
1758 Server-Sent-Events as output.
1760 * The SysV init script compatibility logic will now
1761 heuristically determine whether a script supports the
1762 "reload" verb, and only then make this available as
1765 * "systemctl status --follow" has been removed, use "journalctl
1768 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
1769 have been removed since they are hardly useful to be
1772 * And I'd like to take the opportunity to specifically mention
1773 Zbigniew for his great contributions. Zbigniew, you rock!
1775 Contributions from: Andrew Eikum, Christian Hesse, Colin
1776 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
1777 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
1778 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
1779 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
1780 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
1781 Jędrzejewski-Szmek, Сковорода Никита Андреевич
1785 * If /etc/vconsole.conf is non-existent or empty we will no
1786 longer load any console font or key map at boot by
1787 default. Instead the kernel defaults will be left
1788 intact. This is definitely the right thing to do, as no
1789 configuration should mean no configuration, and hard-coding
1790 font names that are different on all archs is probably a bad
1791 idea. Also, the kernel default key map and font should be
1792 good enough for most cases anyway, and mostly identical to
1793 the userspace fonts/key maps we previously overloaded them
1794 with. If distributions want to continue to default to a
1795 non-kernel font or key map they should ship a default
1796 /etc/vconsole.conf with the appropriate contents.
1798 Contributions from: Colin Walters, Daniel J Walsh, Dave
1799 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
1800 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1804 * journalctl gained a new --cursor= switch to show entries
1805 starting from the specified location in the journal.
1807 * We now enforce a size limit on journal entry fields exported
1808 with "-o json" in journalctl. Fields larger than 4K will be
1809 assigned null. This can be turned off with --all.
1811 * An (optional) journal gateway daemon is now available as
1812 "systemd-journal-gatewayd.service". This service provides
1813 access to the journal via HTTP and JSON. This functionality
1814 will be used to implement live log synchronization in both
1815 pull and push modes, but has various other users too, such
1816 as easy log access for debugging of embedded devices. Right
1817 now it is already useful to retrieve the journal via HTTP:
1819 # systemctl start systemd-journal-gatewayd.service
1820 # wget http://localhost:19531/entries
1822 This will download the journal contents in a
1823 /var/log/messages compatible format. The same as JSON:
1825 # curl -H"Accept: application/json" http://localhost:19531/entries
1827 This service is also accessible via a web browser where a
1828 single static HTML5 app is served that uses the JSON logic
1829 to enable the user to do some basic browsing of the
1830 journal. This will be extended later on. Here's an example
1831 screenshot of this app in its current state:
1833 http://0pointer.de/public/journal-gatewayd
1835 Contributions from: Kay Sievers, Lennart Poettering, Robert
1836 Milasan, Tom Gundersen
1840 * The bash completion logic is now available for journalctl
1843 * We don't mount the "cpuset" controller anymore together with
1844 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
1845 started if no parameters are assigned to it. "cpuset" hence
1846 broke code that assumed it it could create "cpu" groups and
1849 * journalctl -f will now subscribe to terminal size changes,
1850 and line break accordingly.
1852 Contributions from: Dave Reisner, Kay Sievers, Lennart
1853 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
1857 * nspawn will now create a symlink /etc/localtime in the
1858 container environment, copying the host's timezone
1859 setting. Previously this has been done via a bind mount, but
1860 since symlinks cannot be bind mounted this has now been
1861 changed to create/update the appropriate symlink.
1863 * journalctl -n's line number argument is now optional, and
1864 will default to 10 if omitted.
1866 * journald will now log the maximum size the journal files may
1867 take up on disk. This is particularly useful if the default
1868 built-in logic of determining this parameter from the file
1869 system size is used. Use "systemctl status
1870 systemd-journald.service" to see this information.
1872 * The multi-seat X wrapper tool has been stripped down. As X
1873 is now capable of enumerating graphics devices via udev in a
1874 seat-aware way the wrapper is not strictly necessary
1875 anymore. A stripped down temporary stop-gap is still shipped
1876 until the upstream display managers have been updated to
1877 fully support the new X logic. Expect this wrapper to be
1878 removed entirely in one of the next releases.
1880 * HandleSleepKey= in logind.conf has been split up into
1881 HandleSuspendKey= and HandleHibernateKey=. The old setting
1882 is not available anymore. X11 and the kernel are
1883 distuingishing between these keys and we should too. This
1884 also means the inhibition lock for these keys has been split
1887 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
1888 Poettering, Lukas Nykryn, Václav Pavlín
1892 * Whenever a unit changes state we'll now log this to the
1893 journal and show along the unit's own log output in
1896 * ConditionPathIsMountPoint= can now properly detect bind
1897 mount points too. (Previously, a bind mount of one file
1898 system to another place in the same file system could not be
1899 detected as mount, since they shared struct stat's st_dev
1902 * We will now mount the cgroup controllers cpu, cpuacct,
1903 cpuset and the controllers net_cls, net_prio together by
1906 * nspawn containers will now have a virtualized boot
1907 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
1908 over with a randomized ID at container initialization). This
1909 has the effect of making "journalctl -b" do the right thing
1912 * The JSON output journal serialization has been updated not
1913 to generate "endless" list objects anymore, but rather one
1914 JSON object per line. This is more in line how most JSON
1915 parsers expect JSON objects. The new output mode
1916 "json-pretty" has been added to provide similar output, but
1917 neatly aligned for readability by humans.
1919 * We dropped all explicit sync() invocations in the shutdown
1920 code. The kernel does this implicitly anyway in the kernel
1921 reboot() syscall. halt(8)'s -n option is now a compatibility
1924 * We now support virtualized reboot() in containers, as
1925 supported by newer kernels. We will fall back to exit() if
1926 CAP_SYS_REBOOT is not available to the container. Also,
1927 nspawn makes use of this now and will actually reboot the
1928 container if the containerized OS asks for that.
1930 * journalctl will only show local log output by default
1931 now. Use --merge (-m) to show remote log output, too.
1933 * libsystemd-journal gained the new sd_journal_get_usage()
1934 call to determine the current disk usage of all journal
1935 files. This is exposed in the new "journalctl --disk-usage"
1938 * journald gained a new configuration setting SplitMode= in
1939 journald.conf which may be used to control how user journals
1940 are split off. See journald.conf(5) for details.
1942 * A new condition type ConditionFileNotEmpty= has been added.
1944 * tmpfiles' "w" lines now support file globbing, to write
1945 multiple files at once.
1947 * We added Python bindings for the journal submission
1948 APIs. More Python APIs for a number of selected APIs will
1949 likely follow. Note that we intend to add native bindings
1950 only for the Python language, as we consider it common
1951 enough to deserve bindings shipped within systemd. There are
1952 various projects outside of systemd that provide bindings
1953 for languages such as PHP or Lua.
1955 * Many conditions will now resolve specifiers such as %i. In
1956 addition, PathChanged= and related directives of .path units
1957 now support specifiers as well.
1959 * There's now a new RPM macro definition for the system preset
1962 * journald will now warn if it can't forward a message to the
1963 syslog daemon because it's socket is full.
1965 * timedated will no longer write or process /etc/timezone,
1966 except on Debian. As we do not support late mounted /usr
1967 anymore /etc/localtime always being a symlink is now safe,
1968 and hence the information in /etc/timezone is not necessary
1971 * logind will now always reserve one VT for a text getty (VT6
1972 by default). Previously if more than 6 X sessions where
1973 started they took up all the VTs with auto-spawned gettys,
1974 so that no text gettys were available anymore.
1976 * udev will now automatically inform the btrfs kernel logic
1977 about btrfs RAID components showing up. This should make
1978 simple hotplug based btrfs RAID assembly work.
1980 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
1981 (but not for its children which will stay at the kernel
1982 default). This should allow setups with a lot more listening
1985 * systemd will now always pass the configured timezone to the
1986 kernel at boot. timedated will do the same when the timezone
1989 * logind's inhibition logic has been updated. By default,
1990 logind will now handle the lid switch, the power and sleep
1991 keys all the time, even in graphical sessions. If DEs want
1992 to handle these events on their own they should take the new
1993 handle-power-key, handle-sleep-key and handle-lid-switch
1994 inhibitors during their runtime. A simple way to achiveve
1995 that is to invoke the DE wrapped in an invocation of:
1997 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
1999 * Access to unit operations is now checked via SELinux taking
2000 the unit file label and client process label into account.
2002 * systemd will now notify the administrator in the journal
2003 when he over-mounts a non-empty directory.
2005 * There are new specifiers that are resolved in unit files,
2006 for the host name (%H), the machine ID (%m) and the boot ID
2009 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
2010 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
2011 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
2012 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
2013 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
2014 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2015 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
2019 * Support for reading structured kernel messages from
2020 /dev/kmsg has now been added and is enabled by default.
2022 * Support for reading kernel messages from /proc/kmsg has now
2023 been removed. If you want kernel messages in the journal
2024 make sure to run a recent kernel (>= 3.5) that supports
2025 reading structured messages from /dev/kmsg (see
2026 above). /proc/kmsg is now exclusive property of classic
2027 syslog daemons again.
2029 * The libudev API gained the new
2030 udev_device_new_from_device_id() call.
2032 * The logic for file system namespace (ReadOnlyDirectory=,
2033 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
2034 require pivot_root() anymore. This means fewer temporary
2035 directories are created below /tmp for this feature.
2037 * nspawn containers will now see and receive all submounts
2038 made on the host OS below the root file system of the
2041 * Forward Secure Sealing is now supported for Journal files,
2042 which provide cryptographical sealing of journal files so
2043 that attackers cannot alter log history anymore without this
2044 being detectable. Lennart will soon post a blog story about
2045 this explaining it in more detail.
2047 * There are two new service settings RestartPreventExitStatus=
2048 and SuccessExitStatus= which allow configuration of exit
2049 status (exit code or signal) which will be excepted from the
2050 restart logic, resp. consider successful.
2052 * journalctl gained the new --verify switch that can be used
2053 to check the integrity of the structure of journal files and
2054 (if Forward Secure Sealing is enabled) the contents of
2057 * nspawn containers will now be run with /dev/stdin, /dev/fd/
2058 and similar symlinks pre-created. This makes running shells
2059 as container init process a lot more fun.
2061 * The fstab support can now handle PARTUUID= and PARTLABEL=
2064 * A new ConditionHost= condition has been added to match
2065 against the hostname (with globs) and machine ID. This is
2066 useful for clusters where a single OS image is used to
2067 provision a large number of hosts which shall run slightly
2068 different sets of services.
2070 * Services which hit the restart limit will now be placed in a
2073 Contributions from: Bertram Poettering, Dave Reisner, Huang
2074 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
2075 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
2079 * When running in --user mode systemd will now become a
2080 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
2081 tree a lot more organized.
2083 * A new PartOf= unit dependency type has been introduced that
2084 may be used to group services in a natural way.
2086 * "systemctl enable" may now be used to enable instances of
2089 * journalctl now prints error log levels in red, and
2090 warning/notice log levels in bright white. It also supports
2091 filtering by log level now.
2093 * cgtop gained a new -n switch (similar to top), to configure
2094 the maximum number of iterations to run for. It also gained
2095 -b, to run in batch mode (accepting no input).
2097 * The suffix ".service" may now be omitted on most systemctl
2098 command lines involving service unit names.
2100 * There's a new bus call in logind to lock all sessions, as
2101 well as a loginctl verb for it "lock-sessions".
2103 * libsystemd-logind.so gained a new call sd_journal_perror()
2104 that works similar to libc perror() but logs to the journal
2105 and encodes structured information about the error number.
2107 * /etc/crypttab entries now understand the new keyfile-size=
2110 * shutdown(8) now can send a (configurable) wall message when
2111 a shutdown is cancelled.
2113 * The mount propagation mode for the root file system will now
2114 default to "shared", which is useful to make containers work
2115 nicely out-of-the-box so that they receive new mounts from
2116 the host. This can be undone locally by running "mount
2117 --make-rprivate /" if needed.
2119 * The prefdm.service file has been removed. Distributions
2120 should maintain this unit downstream if they intend to keep
2121 it around. However, we recommend writing normal unit files
2122 for display managers instead.
2124 * Since systemd is a crucial part of the OS we will now
2125 default to a number of compiler switches that improve
2126 security (hardening) such as read-only relocations, stack
2127 protection, and suchlike.
2129 * The TimeoutSec= setting for services is now split into
2130 TimeoutStartSec= and TimeoutStopSec= to allow configuration
2131 of individual time outs for the start and the stop phase of
2134 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
2135 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
2136 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
2137 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
2138 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
2139 Gundersen, Zbigniew Jędrzejewski-Szmek
2143 * The journal and id128 C APIs are now fully documented as man
2146 * Extra safety checks have been added when transitioning from
2147 the initial RAM disk to the main system to avoid accidental
2150 * /etc/crypttab entries now understand the new keyfile-offset=
2153 * systemctl -t can now be used to filter by unit load state.
2155 * The journal C API gained the new sd_journal_wait() call to
2156 make writing synchronous journal clients easier.
2158 * journalctl gained the new -D switch to show journals from a
2161 * journalctl now displays a special marker between log
2162 messages of two different boots.
2164 * The journal is now explicitly flushed to /var via a service
2165 systemd-journal-flush.service, rather than implicitly simply
2166 by seeing /var/log/journal to be writable.
2168 * journalctl (and the journal C APIs) can now match for much
2169 more complex expressions, with alternatives and
2172 * When transitioning from the initial RAM disk to the main
2173 system we will now kill all processes in a killing spree to
2174 ensure no processes stay around by accident.
2176 * Three new specifiers may be used in unit files: %u, %h, %s
2177 resolve to the user name, user home directory resp. user
2178 shell. This is useful for running systemd user instances.
2180 * We now automatically rotate journal files if their data
2181 object hash table gets a fill level > 75%. We also size the
2182 hash table based on the configured maximum file size. This
2183 together should lower hash collisions drastically and thus
2184 speed things up a bit.
2186 * journalctl gained the new "--header" switch to introspect
2187 header data of journal files.
2189 * A new setting SystemCallFilters= has been added to services
2190 which may be used to apply blacklists or whitelists to
2191 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
2193 * nspawn gained a new --link-journal= switch (and quicker: -j)
2194 to link the container journal with the host. This makes it
2195 very easy to centralize log viewing on the host for all
2196 guests while still keeping the journal files separated.
2198 * Many bugfixes and optimizations
2200 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
2201 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
2202 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
2207 * Several tools now understand kernel command line arguments,
2208 which are only read when run in an initial RAM disk. They
2209 usually follow closely their normal counterparts, but are
2212 * There's a new tool to analyze the readahead files that are
2213 automatically generated at boot. Use:
2215 /usr/lib/systemd/systemd-readahead analyze /.readahead
2217 * We now provide an early debug shell on tty9 if this enabled. Use:
2219 systemctl enable debug-shell.service
2221 * All plymouth related units have been moved into the Plymouth
2222 package. Please make sure to upgrade your Plymouth version
2225 * systemd-tmpfiles now supports getting passed the basename of
2226 a configuration file only, in which case it will look for it
2227 in all appropriate directories automatically.
2229 * udevadm info now takes a /dev or /sys path as argument, and
2230 does the right thing. Example:
2232 udevadm info /dev/sda
2233 udevadm info /sys/class/block/sda
2235 * systemctl now prints a warning if a unit is stopped but a
2236 unit that might trigger it continues to run. Example: a
2237 service is stopped but the socket that activates it is left
2240 * "systemctl status" will now mention if the log output was
2241 shortened due to rotation since a service has been started.
2243 * The journal API now exposes functions to determine the
2244 "cutoff" times due to rotation.
2246 * journald now understands SIGUSR1 and SIGUSR2 for triggering
2247 immediately flushing of runtime logs to /var if possible,
2248 resp. for triggering immediate rotation of the journal
2251 * It is now considered an error if a service is attempted to
2252 be stopped that is not loaded.
2254 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
2256 * systemd-analyze now supports Python 3
2258 * tmpfiles now supports cleaning up directories via aging
2259 where the first level dirs are always kept around but
2260 directories beneath it automatically aged. This is enabled
2261 by prefixing the age field with '~'.
2263 * Seat objects now expose CanGraphical, CanTTY properties
2264 which is required to deal with very fast bootups where the
2265 display manager might be running before the graphics drivers
2266 completed initialization.
2268 * Seat objects now expose a State property.
2270 * We now include RPM macros for service enabling/disabling
2271 based on the preset logic. We recommend RPM based
2272 distributions to make use of these macros if possible. This
2273 makes it simpler to reuse RPM spec files across
2276 * We now make sure that the collected systemd unit name is
2277 always valid when services log to the journal via
2280 * There's a new man page kernel-command-line(7) detailing all
2281 command line options we understand.
2283 * The fstab generator may now be disabled at boot by passing
2284 fstab=0 on the kernel command line.
2286 * A new kernel command line option modules-load= is now understood
2287 to load a specific kernel module statically, early at boot.
2289 * Unit names specified on the systemctl command line are now
2290 automatically escaped as needed. Also, if file system or
2291 device paths are specified they are automatically turned
2292 into the appropriate mount or device unit names. Example:
2294 systemctl status /home
2295 systemctl status /dev/sda
2297 * The SysVConsole= configuration option has been removed from
2298 system.conf parsing.
2300 * The SysV search path is no longer exported on the D-Bus
2303 * The Names= option is been removed from unit file parsing.
2305 * There's a new man page bootup(7) detailing the boot process.
2307 * Every unit and every generator we ship with systemd now
2308 comes with full documentation. The self-explanatory boot is
2311 * A couple of services gained "systemd-" prefixes in their
2312 name if they wrap systemd code, rather than only external
2313 code. Among them fsck@.service which is now
2314 systemd-fsck@.service.
2316 * The HaveWatchdog property has been removed from the D-Bus
2319 * systemd.confirm_spawn= on the kernel command line should now
2322 * There's a new man page crypttab(5) which details all options
2323 we actually understand.
2325 * systemd-nspawn gained a new --capability= switch to pass
2326 additional capabilities to the container.
2328 * timedated will now read known NTP implementation unit names
2329 from /usr/lib/systemd/ntp-units.d/*.list,
2330 systemd-timedated-ntp.target has been removed.
2332 * journalctl gained a new switch "-b" that lists log data of
2333 the current boot only.
2335 * The notify socket is in the abstract namespace again, in
2336 order to support daemons which chroot() at start-up.
2338 * There is a new Storage= configuration option for journald
2339 which allows configuration of where log data should go. This
2340 also provides a way to disable journal logging entirely, so
2341 that data collected is only forwarded to the console, the
2342 kernel log buffer or another syslog implementation.
2344 * Many bugfixes and optimizations
2346 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
2347 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
2348 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
2349 Shawn Landden, Tom Gundersen
2353 * "systemctl help <unit>" now shows the man page if one is
2356 * Several new man pages have been added.
2358 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
2359 MaxLevelConsole= can now be specified in
2360 journald.conf. These options allow reducing the amount of
2361 data stored on disk or forwarded by the log level.
2363 * TimerSlackNSec= can now be specified in system.conf for
2364 PID1. This allows system-wide power savings.
2366 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
2367 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
2372 * logind is now capable of (optionally) handling power and
2373 sleep keys as well as the lid switch.
2375 * journalctl now understands the syntax "journalctl
2376 /usr/bin/avahi-daemon" to get all log output of a specific
2379 * CapabilityBoundingSet= in system.conf now also influences
2380 the capability bound set of usermode helpers of the kernel.
2382 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
2383 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
2384 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
2385 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
2389 * Note that we skipped 139 releases here in order to set the
2390 new version to something that is greater than both udev's
2391 and systemd's most recent version number.
2393 * udev: all udev sources are merged into the systemd source tree now.
2394 All future udev development will happen in the systemd tree. It
2395 is still fully supported to use the udev daemon and tools without
2396 systemd running, like in initramfs or other init systems. Building
2397 udev though, will require the *build* of the systemd tree, but
2398 udev can be properly *run* without systemd.
2400 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
2401 should be used to create dead device nodes as workarounds for broken
2404 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
2405 no longer supported. udev_monitor_new_from_netlink() needs to be
2406 used to subscribe to events.
2408 * udev: when udevd is started by systemd, processes which are left
2409 behind by forking them off of udev rules, are unconditionally cleaned
2410 up and killed now after the event handling has finished. Services or
2411 daemons must be started as systemd services. Services can be
2412 pulled-in by udev to get started, but they can no longer be directly
2413 forked by udev rules.
2415 * udev: the daemon binary is called systemd-udevd now and installed
2416 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
2417 to adapt to that, create symlink, or rename the binary after building
2420 * libudev no longer provides these symbols:
2421 udev_monitor_from_socket()
2422 udev_queue_get_failed_list_entry()
2423 udev_get_{dev,sys,run}_path()
2424 The versions number was bumped and symbol versioning introduced.
2426 * systemd-loginctl and systemd-journalctl have been renamed
2427 to loginctl and journalctl to match systemctl.
2429 * The config files: /etc/systemd/systemd-logind.conf and
2430 /etc/systemd/systemd-journald.conf have been renamed to
2431 logind.conf and journald.conf. Package updates should rename
2432 the files to the new names on upgrade.
2434 * For almost all files the license is now LGPL2.1+, changed
2435 from the previous GPL2.0+. Exceptions are some minor stuff
2436 of udev (which will be changed to LGPL2.1 eventually, too),
2437 and the MIT licensed sd-daemon.[ch] library that is suitable
2438 to be used as drop-in files.
2440 * systemd and logind now handle system sleep states, in
2441 particular suspending and hibernating.
2443 * logind now implements a sleep/shutdown/idle inhibiting logic
2444 suitable for a variety of uses. Soonishly Lennart will blog
2445 about this in more detail.
2447 * var-run.mount and var-lock.mount are no longer provided
2448 (which prevously bind mounted these directories to their new
2449 places). Distributions which have not converted these
2450 directories to symlinks should consider stealing these files
2451 from git history and add them downstream.
2453 * We introduced the Documentation= field for units and added
2454 this to all our shipped units. This is useful to make it
2455 easier to explore the boot and the purpose of the various
2458 * All smaller setup units (such as
2459 systemd-vconsole-setup.service) now detect properly if they
2460 are run in a container and are skipped when
2461 appropriate. This guarantees an entirely noise-free boot in
2462 Linux container environments such as systemd-nspawn.
2464 * A framework for implementing offline system updates is now
2465 integrated, for details see:
2466 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
2468 * A new service type Type=idle is available now which helps us
2469 avoiding ugly interleaving of getty output and boot status
2472 * There's now a system-wide CapabilityBoundingSet= option to
2473 globally reduce the set of capabilities for the
2474 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
2475 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
2476 even CAP_NET_ADMIN system-wide for secure systems.
2478 * There are now system-wide DefaultLimitXXX= options to
2479 globally change the defaults of the various resource limits
2480 for all units started by PID 1.
2482 * Harald Hoyer's systemd test suite has been integrated into
2483 systemd which allows easy testing of systemd builds in qemu
2484 and nspawn. (This is really awesome! Ask us for details!)
2486 * The fstab parser is now implemented as generator, not inside
2489 * systemctl will now warn you if .mount units generated from
2490 /etc/fstab are out of date due to changes in fstab that
2491 haven't been read by systemd yet.
2493 * systemd is now suitable for usage in initrds. Dracut has
2494 already been updated to make use of this. With this in place
2495 initrds get a slight bit faster but primarily are much
2496 easier to introspect and debug since "systemctl status" in
2497 the host system can be used to introspect initrd services,
2498 and the journal from the initrd is kept around too.
2500 * systemd-delta has been added, a tool to explore differences
2501 between user/admin configuration and vendor defaults.
2503 * PrivateTmp= now affects both /tmp and /var/tmp.
2505 * Boot time status messages are now much prettier and feature
2506 proper english language. Booting up systemd has never been
2509 * Read-ahead pack files now include the inode number of all
2510 files to pre-cache. When the inode changes the pre-caching
2511 is not attempted. This should be nicer to deal with updated
2512 packages which might result in changes of read-ahead
2515 * We now temporaritly lower the kernel's read_ahead_kb variable
2516 when collecting read-ahead data to ensure the kernel's
2517 built-in read-ahead does not add noise to our measurements
2518 of necessary blocks to pre-cache.
2520 * There's now RequiresMountsFor= to add automatic dependencies
2521 for all mounts necessary for a specific file system path.
2523 * MountAuto= and SwapAuto= have been removed from
2524 system.conf. Mounting file systems at boot has to take place
2527 * nspawn now learned a new switch --uuid= to set the machine
2528 ID on the command line.
2530 * nspawn now learned the -b switch to automatically search
2533 * vt102 is now the default TERM for serial TTYs, upgraded from
2536 * systemd-logind now works on VT-less systems.
2538 * The build tree has been reorganized. The individual
2539 components now have directories of their own.
2541 * A new condition type ConditionPathIsReadWrite= is now available.
2543 * nspawn learned the new -C switch to create cgroups for the
2544 container in other hierarchies.
2546 * We now have support for hardware watchdogs, configurable in
2549 * The scheduled shutdown logic now has a public API.
2551 * We now mount /tmp as tmpfs by default, but this can be
2552 masked and /etc/fstab can override it.
2554 * Since udisks doesn't make use of /media anymore we are not
2555 mounting a tmpfs on it anymore.
2557 * journalctl gained a new --local switch to only interleave
2558 locally generated journal files.
2560 * We can now load the IMA policy at boot automatically.
2562 * The GTK tools have been split off into a systemd-ui.
2564 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
2565 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
2566 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
2567 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
2568 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
2569 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
2570 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
2571 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
2572 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
2577 * This is mostly a bugfix release
2579 * Support optional initialization of the machine ID from the
2580 KVM or container configured UUID.
2582 * Support immediate reboots with "systemctl reboot -ff"
2584 * Show /etc/os-release data in systemd-analyze output
2586 * Many bugfixes for the journal, including endianness fixes and
2587 ensuring that disk space enforcement works
2589 * sd-login.h is C++ comptaible again
2591 * Extend the /etc/os-release format on request of the Debian
2594 * We now refuse non-UTF8 strings used in various configuration
2595 and unit files. This is done to ensure we don't pass invalid
2596 data over D-Bus or expose it elsewhere.
2598 * Register Mimo USB Screens as suitable for automatic seat
2601 * Read SELinux client context from journal clients in a race
2604 * Reorder configuration file lookup order. /etc now always
2605 overrides /run in order to allow the administrator to always
2606 and unconditionally override vendor supplied or
2607 automatically generated data.
2609 * The various user visible bits of the journal now have man
2610 pages. We still lack man pages for the journal API calls
2613 * We now ship all man pages in HTML format again in the
2616 Contributions from: Dave Reisner, Dirk Eibach, Frederic
2617 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
2618 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
2623 * This is mostly a bugfix release
2625 * systems lacking /etc/os-release are no longer supported.
2627 * Various functionality updates to libsystemd-login.so
2629 * Track class of PAM logins to distuingish greeters from
2632 Contributions from: Kay Sievers, Lennart Poettering, Michael
2637 * This is an important bugfix release for v41.
2639 * Building man pages is now optional which should be useful
2640 for those building systemd from git but unwilling to install
2643 * Watchdog support for supervising services is now usable. In
2644 a future release support for hardware watchdogs
2645 (i.e. /dev/watchdog) will be added building on this.
2647 * Service start rate limiting is now configurable and can be
2648 turned off per service. When a start rate limit is hit a
2649 reboot can automatically be triggered.
2651 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
2653 Contributions from: Benjamin Franzke, Bill Nottingham,
2654 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
2655 Schmidt, Michał Górny, Piotr Drąg
2659 * The systemd binary is installed /usr/lib/systemd/systemd now;
2660 An existing /sbin/init symlink needs to be adapted with the
2663 * The code that loads kernel modules has been ported to invoke
2664 libkmod directly, instead of modprobe. This means we do not
2665 support systems with module-init-tools anymore.
2667 * Watchdog support is now already useful, but still not
2670 * A new kernel command line option systemd.setenv= is
2671 understood to set system wide environment variables
2672 dynamically at boot.
2674 * We now limit the set of capabilities of systemd-journald.
2676 * We now set SIGPIPE to ignore by default, since it only is
2677 useful in shell pipelines, and has little use in general
2678 code. This can be disabled with IgnoreSIPIPE=no in unit
2681 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
2682 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
2687 * This is mostly a bugfix release
2689 * We now expose the reason why a service failed in the
2690 "Result" D-Bus property.
2692 * Rudimentary service watchdog support (will be completed over
2693 the next few releases.)
2695 * When systemd forks off in order execute some service we will
2696 now immediately changes its argv[0] to reflect which process
2697 it will execute. This is useful to minimize the time window
2698 with a generic argv[0], which makes bootcharts more useful
2700 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
2701 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
2702 Mike Kazantsev, Ray Strode
2706 * This is mostly a test release, but incorporates many
2709 * New systemd-cgtop tool to show control groups by their
2712 * Linking against libacl for ACLs is optional again. If
2713 disabled, support tracking device access for active logins
2714 goes becomes unavailable, and so does access to the user
2715 journals by the respective users.
2717 * If a group "adm" exists, journal files are automatically
2718 owned by them, thus allow members of this group full access
2719 to the system journal as well as all user journals.
2721 * The journal now stores the SELinux context of the logging
2722 client for all entries.
2724 * Add C++ inclusion guards to all public headers
2726 * New output mode "cat" in the journal to print only text
2727 messages, without any meta data like date or time.
2729 * Include tiny X server wrapper as a temporary stop-gap to
2730 teach XOrg udev display enumeration. This is used by display
2731 managers such as gdm, and will go away as soon as XOrg
2732 learned native udev hotplugging for display devices.
2734 * Add new systemd-cat tool for executing arbitrary programs
2735 with STDERR/STDOUT connected to the journal. Can also act as
2736 BSD logger replacement, and does so by default.
2738 * Optionally store all locally generated coredumps in the
2739 journal along with meta data.
2741 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
2742 writing short strings to files (for usage for /sys), and for
2743 creating symlinks, character and block device nodes.
2745 * New unit file option ControlGroupPersistent= to make cgroups
2746 persistent, following the mechanisms outlined in
2747 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
2749 * Support multiple local RTCs in a sane way
2751 * No longer monopolize IO when replaying readahead data on
2752 rotating disks, since we might starve non-file-system IO to
2753 death, since fanotify() will not see accesses done by blkid,
2756 * Don't show kernel threads in systemd-cgls anymore, unless
2757 requested with new -k switch.
2759 Contributions from: Dan Horák, Kay Sievers, Lennart
2760 Poettering, Michal Schmidt
2764 * This is mostly a test release, but incorporates many
2767 * The git repository moved to:
2768 git://anongit.freedesktop.org/systemd/systemd
2769 ssh://git.freedesktop.org/git/systemd/systemd
2771 * First release with the journal
2772 http://0pointer.de/blog/projects/the-journal.html
2774 * The journal replaces both systemd-kmsg-syslogd and
2775 systemd-stdout-bridge.
2777 * New sd_pid_get_unit() API call in libsystemd-logind
2779 * Many systemadm clean-ups
2781 * Introduce remote-fs-pre.target which is ordered before all
2782 remote mounts and may be used to start services before all
2785 * Added Mageia support
2787 * Add bash completion for systemd-loginctl
2789 * Actively monitor PID file creation for daemons which exit in
2790 the parent process before having finished writing the PID
2791 file in the daemon process. Daemons which do this need to be
2792 fixed (i.e. PID file creation must have finished before the
2793 parent exits), but we now react a bit more gracefully to them.
2795 * Add colourful boot output, mimicking the well-known output
2796 of existing distributions.
2798 * New option PassCredentials= for socket units, for
2799 compatibility with a recent kernel ABI breakage.
2801 * /etc/rc.local is now hooked in via a generator binary, and
2802 thus will no longer act as synchronization point during
2805 * systemctl list-unit-files now supports --root=.
2807 * systemd-tmpfiles now understands two new commands: z, Z for
2808 relabelling files according to the SELinux database. This is
2809 useful to apply SELinux labels to specific files in /sys,
2812 * Output of SysV services is now forwarded to both the console
2813 and the journal by default, not only just the console.
2815 * New man pages for all APIs from libsystemd-login.
2817 * The build tree got reorganized and a the build system is a
2818 lot more modular allowing embedded setups to specifically
2819 select the components of systemd they are interested in.
2821 * Support for Linux systems lacking the kernel VT subsystem is
2824 * configure's --with-rootdir= got renamed to
2825 --with-rootprefix= to follow the naming used by udev and
2828 * Unless specified otherwise we'll now install to /usr instead
2829 of /usr/local by default.
2831 * Processes with '@' in argv[0][0] are now excluded from the
2832 final shut-down killing spree, following the logic explained
2834 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
2836 * All processes remaining in a service cgroup when we enter
2837 the START or START_PRE states are now killed with
2838 SIGKILL. That means it is no longer possible to spawn
2839 background processes from ExecStart= lines (which was never
2840 supported anyway, and bad style).
2842 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
2843 reloading of units together.
2845 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
2846 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
2847 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
2848 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
2849 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek