1 systemd System and Service Manager
5 (This changes list is very incomplete, and the release is
8 * Introduce a new API "sd-hwdb.h" for querying the hardware
9 metadata database. With this minimal interface one can query
10 and enumerate the udev hwdb, decoupled from the old libudev
11 library. libudev's interface for this is now only a wrapper
12 around sd-hwdb. A new tool systemd-hwdb has been added to
13 interface with and update the database.
15 * When any of systemd's tools copies files (for example due to
16 tmpfiles' C lines) a btrfs reflink will attempted first,
17 before bytewise copying is done.
19 * systemd-nspawn gained a new --ephemeral switch. When
20 specified a btrfs snapshot is taken of the container's root
21 directory, and immediately removed when the container
22 terminates again. Thus, a container can be started whose
23 changes never alter the container's root directory, and are
24 lost on container termination. This switch can also be used
25 for starting a container off the root file system of the
26 host without affecting the host OS. This switch is only
27 available on btrfs file systems.
29 * systemd-nspawn gained a new --template= switch. It takes the
30 path to a container tree to use as template for the tree
31 specified via --directory=, should that director be
32 missing. This allows instantiating containers dynamically,
33 on first run. This switch is only available on btrfs file
36 * When a .mount unit refers to a mount point on which multiple
37 mounts are stacked, and the .mount unit is stopped all of
38 the stacked mount points will now be unmounted until no
41 * systemd now has an explicit notion of supported and
42 unsupported unit types. Jobs enqueued for unsupported unit
43 types will now fail with an "unsupported" error code. More
44 specifically .swap, .automount and .device units are not
45 supported in containers, .busname units are not supported on
46 non-kdbus systems. .swap and .automount are also not
47 supported if their respective kernel compile time options
50 * machinectl gained support for two new "copy-from" and
51 "copy-to" commands for copying files from a running
52 container to the host or vice versa.
54 * machinectl gained support for a new "bind" command to bind
55 mount host directories into local containers. This is
56 currently only supported for nspawn containers.
58 * networkd gained support for configuring bridge forwarding
59 database entries (fdb) from .network files.
61 * A new tiny daemon "systemd-importd" has been added that can
62 download container images in tar, raw, qcow2 or dkr formats,
63 and make them available locally in /var/lib/machines, so
64 that they can run as nspawn containers. The daemon can GPG
65 verify the downloads (not supported for dkr, since it has no
66 provisions for verifying downloads). It will transparently
67 decompress bz2, xz, gzip compressed downloads if necessary,
68 and restore sparse files on disk. The daemon uses privilege
69 separation to ensure the actual download logic runs with
70 fewer privileges than the deamon itself. machinectl has
71 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
72 make the functionality of importd available to the
73 user. With this in place the Fedora and Ubuntu "Cloud"
74 images can be downloaded and booted as containers unmodified
75 (the Fedora images lack the appropriate GPG signature files
76 currently, so they cannot be verified, but this will change
77 soon, hopefully). Note that downloading images is currently
78 only fully supported on btrfs.
80 * machinectl is now able to list container images found in
81 /var/lib/machines, along with some metadata about sizes of
82 disk and similar. If the directory is located on btrfs and
83 quota is enabled, this includes quota display. A new command
84 "image-status" has been added that shows additional
85 information about images.
87 * machinectl is now able to clone container images
88 efficiently, if the underlying file system (btrfs) supports
89 it, with the new "machinectl list-images" command. It also
90 gained commands for renaming and removing images, as well as
91 marking them read-only or read-write (supported also on
94 * networkd gained support for collecting LLDP network
95 announcements, from hardware that supports this. This is
96 shown in networkctl output.
98 * systemd-run gained support for a new -t (--pty) switch for
99 invoking a binary on a pty whose input and output is
100 connected to the invoking terminal. This allows executing
101 processes as system services while interactively
102 communicating with them via the terminal. Most interestingly
103 this is supported across container boundaries. Invoking
104 "systemd-run -t /bin/bash" is an alternative to running a
105 full login session, the difference being that the former
106 will not register a session, nor go through the PAM session
109 * tmpfiles gained support for a new "v" line type for creating
110 btrfs subvolumes. If the underlying file system is a legacy
111 file system, this automatically degrades to creating a
112 normal directory. Among others /var/lib/machines is now
113 created like this at boot, should it be missing.
115 * The directory /var/lib/containers/ has been deprecated and
116 been replaced by /var/lib/machines. The term "machines" has
117 been used in the systemd context as generic term for both
118 VMs and containers, and hence appears more appropriate for
119 this, as the directory can also contain raw images bootable
122 * systemd-nspawn when invoked with -M but without --directory=
123 or --image= is now capable of searching for the container
124 root directory, subvolume or disk image automatically, in
125 /var/lib/machines. systemd-nspawn@.service has been updated
126 to make use of this, thus allowing it to be used for raw
129 * A new machines.target unit has been introduced that is
130 supposed to group all containers/VMs invoked as services on
131 the system. systemd-nspawn@.service has been updated to
134 * machinectl gained a new "start" command, for invoking a
135 container as a service. "machinectl start foo" is mostly
136 equivalent to "systemctl start systemd-nspawn@foo.service",
137 but handles escaping in a nicer way.
139 * systemd-nspawn will now mount most of the cgroupfs tree
140 read-only into each container, with the exception of the
141 container's own subtree in the name=systemd hierarchy.
143 * journald now sets the special FS_NOCOW file flag for its
144 journal files. This should improve performance on btrfs, by
145 avoiding heavy fragmentation when journald's write-pattern
146 is used on COW file systems. It degrades btrfs' data
147 integrity guarantees for the files to the same levels as for
148 ext3/ext4 however. This should be OK though as journald does
149 its own data integrity checks and all its objects are
150 checksummed on disk. Also, journald should handle btrfs disk
151 full events a lot more gracefully now, by processing SIGBUS
152 errors, and not relying on fallocate() anymore.
154 * When journald detects that journal files it is writing to
155 have been deleted it will immediately start new journal
158 * systemd now provides a way to store file descriptors
159 per-service in PID 1.This is useful for daemons to ensure
160 that fds they require are not lost during a daemon
161 restart. The fds are passed to the deamon on the next
162 invocation in the same way socket activation fds are
163 passed. This is now used by journald to ensure that the
164 various sockets connected to all the system's stdout/stderr
165 are not lost when journald is restarted. File descriptors
166 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
167 an extension to sd_notify(). Note that a limit is enforced
168 on the number of fds a service can store in PID 1, and it
169 defaults to 0, so that no fds may be stored, unless this is
170 explicitly turned on.
172 * The default TERM variable to use for units connected to a
173 terminal, when no other value is explicitly is set is now
174 vt220 rather than vt102. This should be fairly safe still,
175 but allows PgUp/PgDn work.
177 * The /etc/crypttab option header= as known from Debian is now
180 * "loginctl user-status" and "loginctl session-status" will
181 now show the last 10 lines of log messages of the
182 user/session following the status output. Similar,
183 "machinectl status" will show the last 10 log lines
184 associated with a virtual machine or container
185 service. (Note that this is usually not the log messages
186 done in the VM/container itself, but simply what the
187 container manager logs. For nspawn this includes all console
190 * "loginctl session-status" without further argument will now
191 show the status of the session of the caller. Similar,
192 "lock-session", "unlock-session", "activate",
193 "enable-linger", "disable-linger" may now be called without
194 session/user parameter in which case they apply to the
195 caller's session/user.
197 * An X11 session scriptlet is now shipped that uploads
198 $DISPLAY and $XAUTHORITY into the environment of the systemd
199 --user daemon if a session begins. This should improve
200 compatibility with X11 enabled applications run as systemd
203 * Generators are now subject to masking via /etc and /run, the
204 same way as unit files.
206 * networkd .network files gained support for configuring
207 per-link IPv4/IPv6 packet forwarding as well as IPv4
208 masquerading. This is by default turned on for veth links to
209 containers, as registered by systemd-nspawn. This means that
210 nspawn containers run with --network-veth will now get
211 automatic routed access to the host's networks without any
212 further configuration or setup, as long as networkd runs on
215 * systemd-nspawn gained the --port= (-p) switch to expose TCP
216 or UDP posts of a container on the host. With this in place
217 it is possible to run containers with private veth links
218 (--network-veth), and have their functionality exposed on
219 the host as if their services were running directly on the
222 * systemd-nspawn's --network-beth switch now gained a short
223 version "-n", since with the changes above it is now truly
224 useful out-of-the-box. The systemd-nspawn@.service has been
225 updated to make use of it too by default.
227 * systemd-nspawn will now maintain a per-image R/W lock, to
228 ensure that the same image is not started more than once
229 writable. (It's OK to run an image multiple times
230 simultaneously in read-only mode.)
232 * systemd-nspawn's --image= option is now capable of
233 dissecting and booting MBR and GPT disk images that contain
234 only a single active Linux partition. Previously it
235 supported only GPT disk images with proper GPT type
236 IDs. This allows running cloud images from major
237 distributions directly with systemd-nspawn, without
240 * In addition to collecting mouse dpi data in the udev
241 hardware database, there's now support for collecting angle
242 information for mouse scroll wheels. The database is
243 supposed to guarantee similar scrolling behaviour on mice
244 that it knows about. There's also support for collecting
245 information about Touchpad types.
247 * udev's input_id built-in will now also collect touch screen
248 dimension data and attach it to probed devices.
250 * /etc/os-release gained support for a Distribution Privacy
253 * networkd gained support for creating "ipvlan", "gretap",
254 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
256 * systemd-tmpfiles gained support for "a" lines for setting
259 * systemd-nspawn will now mount /tmp in the container to
260 tmpfs, automatically.
262 * systemd now exposes the memory.usage_in_bytes cgroup
263 attribute and shows it for each service in the "systemctl
264 status" output, if available.
266 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
267 immediate reboot is triggered. This useful if shutdown is
268 hung and is unable to complete, to expedite the
269 operation. Note that this kind of reboot will still unmount
270 all file systems, and hence should not result in fsck being
273 * A .device unit for an optical block device will now be
274 considered active only when a medium is in the drive. Also,
275 mount units are now bound to their backing devices thus
276 triggering automatic unmounting when devices become
277 unavailable. With this in place systemd will now
278 automatically unmount left-over mounts when a CD-ROM is
279 ejected or an USB stick is yanked from the system.
281 * networkd-wait-online now has support for waiting for
282 specific interfaces only (with globbing), and for giving up
283 after a configurable timeout.
285 * networkd now exits when idle. It will be automatically
286 restarted as soon as interfaces show up, are removed or
287 change state. networkd will stay around as long as there is
288 at least one DHCP state machine or similar around, that keep
291 * networkd may now configure IPv6 link-local addressing in
292 addition to IPv4 link-local addressing.
294 * The IPv6 "token" for use in SLAAC may now be configured for
295 each .network interface in networkd.
297 * Routes configured with networkd may now be assigned a scope
300 * networkd's [Match] sections now support globbing and lists
301 of multiple space-separated matches per item.
303 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
304 Bastien Nocera, Carlos Garnacho, Carlos Morata Castillo, Chris
305 Atkinson, Chris J. Arges, Christian Seiler, Christoph Brill,
306 Colin Guthrie, Colin Walters, Cristian Rodríguez, Daniele
307 Medri, Daniel Mack, Dave Reisner, David Herrmann, Erik
308 Auerswald, Filipe Brandenburger, Frank Theile, Gabor Kelemen,
309 Gabriel de Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov,
310 Jan Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl,
311 Jonas Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers,
312 Ken Werner, Lennart Poettering, Lucas De Marchi, Lukas Nykryn,
313 Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas, Manuel
314 Mendez, Marcel Holtmann, Marc Schmitzer, Marko Myllynen,
315 Martin Pitt, Maxim Mikityanskiy, Michael Biebl, Michael
316 Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
317 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
318 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
319 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
320 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
321 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
322 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
323 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
324 Lindskog, Veres Lajos, Vincent Batts, Wieland Hoffmann,
325 Zbigniew Jędrzejewski-Szmek
327 -- Berlin, 2015-02-??
331 * When querying unit file enablement status (for example via
332 "systemctl is-enabled"), a new state "indirect" is now known
333 which indicates that a unit might not be enabled itself, but
334 another unit listed in its Alias= setting might be.
336 * Similar to the various existing ConditionXYZ= settings for
337 units there are now matching AssertXYZ= settings. While
338 failing conditions cause a unit to be skipped, but its job
339 to succeed, failing assertions declared like this will cause
340 a unit start operation and its job to fail.
342 * hostnamed now knows a new chassis type "embedded".
344 * systemctl gained a new "edit" command. When used on a unit
345 file this allows extending unit files with .d/ drop-in
346 configuration snippets or editing the full file (after
347 copying it from /usr/lib to /etc). This will invoke the
348 user's editor (as configured with $EDITOR), and reload the
349 modified configuration after editing.
351 * "systemctl status" now shows the suggested enablement state
352 for a unit, as declared in the (usually vendor-supplied)
355 * nss-myhostname will now resolve the single-label host name
356 "gateway" to the locally configured default IP routing
357 gateways, ordered by their metrics. This assigns a stable
358 name to the used gateways, regardless which ones are
359 currently configured. Note that the name will only be
360 resolved after all other name sources (if nss-myhostname is
361 configured properly) and should hence not negatively impact
362 systems that use the single-label host name "gateway" in
365 * systemd-inhibit now allows filtering by mode when listing
368 * Scope and service units gained a new "Delegate" boolean
369 property, which when set allows processes running inside the
370 unit to further partition resources. This is primarily
371 useful for systemd user instances as well as container
374 * journald will now pick up audit messages directly from
375 the kernel, and log them like any other log message. The
376 audit fields are split up and fully indexed. This means that
377 journalctl in many ways is now a (nicer!) alternative to
378 ausearch, the traditional audit client. Note that this
379 implements only a minimal audit client, if you want the
380 special audit modes like reboot-on-log-overflow, please use
381 the traditional auditd instead, which can be used in
382 parallel to journald.
384 * The ConditionSecurity= unit file option now understands the
385 special string "audit" to check whether auditing is
388 * journalctl gained two new commands --vacuum-size= and
389 --vacuum-time= to delete old journal files until the
390 remaining ones take up no more the specified size on disk,
391 or are not older than the specified time.
393 * A new, native PPPoE library has been added to sd-network,
394 systemd's library of light-weight networking protocols. This
395 library will be used in a future version of networkd to
396 enable PPPoE communication without an external pppd daemon.
398 * The busctl tool now understands a new "capture" verb that
399 works similar to "monitor", but writes a packet capture
400 trace to STDOUT that can be redirected to a file which is
401 compatible with libcap's capture file format. This can then
402 be loaded in Wireshark and similar tools to inspect bus
405 * The busctl tool now understands a new "tree" verb that shows
406 the object trees of a specific service on the bus, or of all
409 * The busctl tool now understands a new "introspect" verb that
410 shows all interfaces and members of objects on the bus,
411 including their signature and values. This is particularly
412 useful to get more information about bus objects shown by
413 the new "busctl tree" command.
415 * The busctl tool now understands new verbs "call",
416 "set-property" and "get-property" for invoking bus method
417 calls, setting and getting bus object properties in a
420 * busctl gained a new --augment-creds= argument that controls
421 whether the tool shall augment credential information it
422 gets from the bus with data from /proc, in a possibly
425 * nspawn's --link-journal= switch gained two new values
426 "try-guest" and "try-host" that work like "guest" and
427 "host", but do not fail if the host has no persistent
428 journalling enabled. -j is now equivalent to
429 --link-journal=try-guest.
431 * macvlan network devices created by nspawn will now have
432 stable MAC addresses.
434 * A new SmackProcessLabel= unit setting has been added, which
435 controls the SMACK security label processes forked off by
436 the respective unit shall use.
438 * If compiled with --enable-xkbcommon, systemd-localed will
439 verify x11 keymap settings by compiling the given keymap. It
440 will spew out warnings if the compilation fails. This
441 requires libxkbcommon to be installed.
443 * When a coredump is collected a larger number of metadata
444 fields is now collected and included in the journal records
445 created for it. More specifically control group membership,
446 environment variables, memory maps, working directory,
447 chroot directory, /proc/$PID/status, and a list of open file
448 descriptors is now stored in the log entry.
450 * The udev hwdb now contains DPI information for mice. For
453 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
455 * All systemd programs that read standalone configuration
456 files in /etc now also support a corresponding series of
457 .conf.d configuration directories in /etc/, /run/,
458 /usr/local/lib/, /usr/lib/, and (if configured with
459 --enable-split-usr) /lib/. In particular, the following
460 configuration files now have corresponding configuration
461 directories: system.conf user.conf, logind.conf,
462 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
463 resolved.conf, timesyncd.conf, journal-remote.conf, and
464 journal-upload.conf. Note that distributions should use the
465 configuration directories in /usr/lib/; the directories in
466 /etc/ are reserved for the system administrator.
468 * systemd-rfkill will no longer take the rfkill device name
469 into account when storing rfkill state on disk, as the name
470 might be dynamically assigned and not stable. Instead, the
471 ID_PATH udev variable combined with the rfkill type (wlan,
472 bluetooth, ...) is used.
474 * A new service systemd-machine-id-commit.service has been
475 added. When used on systems where /etc is read-only during
476 boot, and /etc/machine-id is not initialized (but an empty
477 file), this service will copy the temporary machine ID
478 created as replacement into /etc after the system is fully
479 booted up. This is useful for systems that are freshly
480 installed with a non-initialized machine ID, but should get
481 a fixed machine ID for subsequent boots.
483 * networkd's .netdev files now provide a large set of
484 configuration parameters for VXLAN devices. Similar, the
485 bridge port cost parameter is now configurable in .network
486 files. There's also new support for configuring IP source
487 routing. networkd .link files gained support for a new
488 OriginalName= match that is useful to match against the
489 original interface name the kernel assigned. .network files
490 may include MTU= and MACAddress= fields for altering the MTU
491 and MAC address while being connected to a specific network
494 * The LUKS logic gained supported for configuring
495 UUID-specific key files. There's also new support for naming
496 LUKS device from the kernel command line, using the new
499 * Timer units may now be transiently created via the bus API
500 (this was previously already available for scope and service
501 units). In addition it is now possible to create multiple
502 transient units at the same time with a single bus call. The
503 "systemd-run" tool has been updated to make use of this for
504 running commands on a specified time, in at(1)-style.
506 * tmpfiles gained support for "t" lines, for assigning
507 extended attributes to files. Among other uses this may be
508 used to assign SMACK labels to files.
510 Contributions from: Alin Rauta, Alison Chaiken, Andrej
511 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
512 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
513 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
514 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
515 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
516 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
517 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
518 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
519 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
520 Michael Biebl, Michael Chapman, Michael Marineau, Michal
521 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
522 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
523 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
524 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
525 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
526 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
527 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
529 -- Berlin, 2014-12-10
533 * journalctl gained the new options -t/--identifier= to match
534 on the syslog identifier (aka "tag"), as well as --utc to
535 show log timestamps in the UTC timezone. journalctl now also
536 accepts -n/--lines=all to disable line capping in a pager.
538 * journalctl gained a new switch, --flush, that synchronously
539 flushes logs from /run/log/journal to /var/log/journal if
540 persistent storage is enabled. systemd-journal-flush.service
541 now waits until the operation is complete.
543 * Services can notify the manager before they start a reload
544 (by sending RELOADING=1) or shutdown (by sending
545 STOPPING=1). This allows the manager to track and show the
546 internal state of daemons and closes a race condition when
547 the process is still running but has closed its D-Bus
550 * Services with Type=oneshot do not have to have any ExecStart
553 * User units are now loaded also from
554 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
555 /run/systemd/user directory that was already previously
556 supported, but is under the control of the user.
558 * Job timeouts (i.e. time-outs on the time a job that is
559 queued stays in the run queue) can now optionally result in
560 immediate reboot or power-off actions (JobTimeoutAction= and
561 JobTimeoutRebootArgument=). This is useful on ".target"
562 units, to limit the maximum time a target remains
563 undispatched in the run queue, and to trigger an emergency
564 operation in such a case. This is now used by default to
565 turn off the system if boot-up (as defined by everything in
566 basic.target) hangs and does not complete for at least
567 15min. Also, if power-off or reboot hang for at least 30min
568 an immediate power-off/reboot operation is triggered. This
569 functionality is particularly useful to increase reliability
570 on embedded devices, but also on laptops which might
571 accidentally get powered on when carried in a backpack and
572 whose boot stays stuck in a hard disk encryption passphrase
575 * systemd-logind can be configured to also handle lid switch
576 events even when the machine is docked or multiple displays
577 are attached (HandleLidSwitchDocked= option).
579 * A helper binary and a service have been added which can be
580 used to resume from hibernation in the initramfs. A
581 generator will parse the resume= option on the kernel
582 command line to trigger resume.
584 * A user console daemon systemd-consoled has been
585 added. Currently, it is a preview, and will so far open a
586 single terminal on each session of the user marked as
587 Desktop=systemd-console.
589 * Route metrics can be specified for DHCP routes added by
592 * The SELinux context of socket-activated services can be set
593 from the information provided by the networking stack
594 (SELinuxContextFromNet= option).
596 * Userspace firmware loading support has been removed and
597 the minimum supported kernel version is thus bumped to 3.7.
599 * Timeout for udev workers has been increased from 1 to 3
600 minutes, but a warning will be printed after 1 minute to
601 help diagnose kernel modules that take a long time to load.
603 * Udev rules can now remove tags on devices with TAG-="foobar".
605 * systemd's readahead implementation has been removed. In many
606 circumstances it didn't give expected benefits even for
607 rotational disk drives and was becoming less relevant in the
608 age of SSDs. As none of the developers has been using
609 rotating media anymore, and nobody stepped up to actively
610 maintain this component of systemd it has now been removed.
612 * Swap units can use Options= to specify discard options.
613 Discard options specified for swaps in /etc/fstab are now
616 * Docker containers are now detected as a separate type of
619 * The Password Agent protocol gained support for queries where
620 the user input is shown, useful e.g. for user names.
621 systemd-ask-password gained a new --echo option to turn that
624 * The default sysctl.d/ snippets will now set:
626 net.core.default_qdisc = fq_codel
628 This selects Fair Queuing Controlled Delay as the default
629 queuing discipline for network interfaces. fq_codel helps
630 fight the network bufferbloat problem. It is believed to be
631 a good default with no tuning required for most workloads.
632 Downstream distributions may override this choice. On 10Gbit
633 servers that do not do forwarding, "fq" may perform better.
634 Systems without a good clocksource should use "pfifo_fast".
636 * If kdbus is enabled during build a new option BusPolicy= is
637 available for service units, that allows locking all service
638 processes into a stricter bus policy, in order to limit
639 access to various bus services, or even hide most of them
640 from the service's view entirely.
642 * networkctl will now show the .network and .link file
643 networkd has applied to a specific interface.
645 * sd-login gained a new API call sd_session_get_desktop() to
646 query which desktop environment has been selected for a
649 * UNIX utmp support is now compile-time optional to support
652 * systemctl gained two new commands "add-wants" and
653 "add-requires" for pulling in units from specific targets
656 * If the word "rescue" is specified on the kernel command line
657 the system will now boot into rescue mode (aka
658 rescue.target), which was previously available only by
659 specifying "1" or "systemd.unit=rescue.target" on the kernel
660 command line. This new kernel command line option nicely
661 mirrors the already existing "emergency" kernel command line
664 * New kernel command line options mount.usr=, mount.usrflags=,
665 mount.usrfstype= have been added that match root=, rootflags=,
666 rootfstype= but allow mounting a specific file system to
669 * The $NOTIFY_SOCKET is now also passed to control processes of
670 services, not only the main process.
672 * This version reenables support for fsck's -l switch. This
673 means at least version v2.25 of util-linux is required for
674 operation, otherwise dead-locks on device nodes may
675 occur. Again: you need to update util-linux to at least
676 v2.25 when updating systemd to v217.
678 * The "multi-seat-x" tool has been removed from systemd, as
679 its functionality has been integrated into X servers 1.16,
680 and the tool is hence redundant. It is recommended to update
681 display managers invoking this tool to simply invoke X
682 directly from now on, again.
684 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
685 message flag has been added for all of systemd's PolicyKit
686 authenticated method calls has been added. In particular
687 this now allows optional interactive authorization via
688 PolicyKit for many of PID1's privileged operations such as
689 unit file enabling and disabling.
691 * "udevadm hwdb --update" learnt a new switch "--usr" for
692 placing the rebuilt hardware database in /usr instead of
693 /etc. When used only hardware database entries stored in
694 /usr will be used, and any user database entries in /etc are
695 ignored. This functionality is useful for vendors to ship a
696 pre-built database on systems where local configuration is
697 unnecessary or unlikely.
699 * Calendar time specifications in .timer units now also
700 understand the strings "semi-annually", "quarterly" and
701 "minutely" as shortcuts (in addition to the preexisting
702 "anually", "hourly", ...).
704 * systemd-tmpfiles will now correctly create files in /dev
705 at boot which are marked for creation only at boot. It is
706 recommended to always create static device nodes with 'c!'
707 and 'b!', so that they are created only at boot and not
708 overwritten at runtime.
710 * When the watchdog logic is used for a service (WatchdogSec=)
711 and the watchdog timeout is hit the service will now be
712 terminated with SIGABRT (instead of just SIGTERM), in order
713 to make sure a proper coredump and backtrace is
714 generated. This ensures that hanging services will result in
715 similar coredump/backtrace behaviour as services that hit a
718 Contributions from: Andreas Henriksson, Andrei Borzenkov,
719 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
720 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
721 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
722 Herrmann, David Sommerseth, David Strauss, Emil Renner
723 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
724 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
725 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
726 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
727 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
728 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
729 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
730 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
731 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
732 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
733 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
734 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
735 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
736 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
737 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
738 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
741 -- Berlin, 2014-10-28
745 * timedated no longer reads NTP implementation unit names from
746 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
747 implementations should add a
749 Conflicts=systemd-timesyncd.service
751 to their unit files to take over and replace systemd's NTP
752 default functionality.
754 * systemd-sysusers gained a new line type "r" for configuring
755 which UID/GID ranges to allocate system users/groups
756 from. Lines of type "u" may now add an additional column
757 that specifies the home directory for the system user to be
758 created. Also, systemd-sysusers may now optionally read user
759 information from STDIN instead of a file. This is useful for
760 invoking it from RPM preinst scriptlets that need to create
761 users before the first RPM file is installed since these
762 files might need to be owned by them. A new
763 %sysusers_create_inline RPM macro has been introduced to do
764 just that. systemd-sysusers now updates the shadow files as
765 well as the user/group databases, which should enhance
766 compatibility with certain tools like grpck.
768 * A number of bus APIs of PID 1 now optionally consult
769 PolicyKit to permit access for otherwise unprivileged
770 clients under certain conditions. Note that this currently
771 doesn't support interactive authentication yet, but this is
772 expected to be added eventually, too.
774 * /etc/machine-info now has new fields for configuring the
775 deployment environment of the machine, as well as the
776 location of the machine. hostnamectl has been updated with
777 new command to update these fields.
779 * systemd-timesyncd has been updated to automatically acquire
780 NTP server information from systemd-networkd, which might
781 have been discovered via DHCP.
783 * systemd-resolved now includes a caching DNS stub resolver
784 and a complete LLMNR name resolution implementation. A new
785 NSS module "nss-resolve" has been added which can be used
786 instead of glibc's own "nss-dns" to resolve hostnames via
787 systemd-resolved. Hostnames, addresses and arbitrary RRs may
788 be resolved via systemd-resolved D-Bus APIs. In contrast to
789 the glibc internal resolver systemd-resolved is aware of
790 multi-homed system, and keeps DNS server and caches separate
791 and per-interface. Queries are sent simultaneously on all
792 interfaces that have DNS servers configured, in order to
793 properly handle VPNs and local LANs which might resolve
794 separate sets of domain names. systemd-resolved may acquire
795 DNS server information from systemd-networkd automatically,
796 which in turn might have discovered them via DHCP. A tool
797 "systemd-resolve-host" has been added that may be used to
798 query the DNS logic in resolved. systemd-resolved implements
799 IDNA and automatically uses IDNA or UTF-8 encoding depending
800 on whether classic DNS or LLMNR is used as transport. In the
801 next releases we intend to add a DNSSEC and mDNS/DNS-SD
802 implementation to systemd-resolved.
804 * A new NSS module nss-mymachines has been added, that
805 automatically resolves the names of all local registered
806 containers to their respective IP addresses.
808 * A new client tool "networkctl" for systemd-networkd has been
809 added. It currently is entirely passive and will query
810 networking configuration from udev, rtnetlink and networkd,
811 and present it to the user in a very friendly
812 way. Eventually, we hope to extend it to become a full
813 control utility for networkd.
815 * .socket units gained a new DeferAcceptSec= setting that
816 controls the kernels' TCP_DEFER_ACCEPT sockopt for
817 TCP. Similar, support for controlling TCP keep-alive
818 settings has been added (KeepAliveTimeSec=,
819 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
820 turning off Nagle's algorithm on TCP has been added
823 * logind learned a new session type "web", for use in projects
824 like Cockpit which register web clients as PAM sessions.
826 * timer units with at least one OnCalendar= setting will now
827 be started only after timer-sync.target has been
828 reached. This way they will not elapse before the system
829 clock has been corrected by a local NTP client or
830 similar. This is particular useful on RTC-less embedded
831 machines, that come up with an invalid system clock.
833 * systemd-nspawn's --network-veth= switch should now result in
834 stable MAC addresses for both the outer and the inner side
837 * systemd-nspawn gained a new --volatile= switch for running
838 container instances with /etc or /var unpopulated.
840 * The kdbus client code has been updated to use the new Linux
841 3.17 memfd subsystem instead of the old kdbus-specific one.
843 * systemd-networkd's DHCP client and server now support
844 FORCERENEW. There are also new configuration options to
845 configure the vendor client identifier and broadcast mode
848 * systemd will no longer inform the kernel about the current
849 timezone, as this is necessarily incorrect and racy as the
850 kernel has no understanding of DST and similar
851 concepts. This hence means FAT timestamps will be always
852 considered UTC, similar to what Android is already
853 doing. Also, when the RTC is configured to the local time
854 (rather than UTC) systemd will never synchronize back to it,
855 as this might confuse Windows at a later boot.
857 * systemd-analyze gained a new command "verify" for offline
858 validation of unit files.
860 * systemd-networkd gained support for a couple of additional
861 settings for bonding networking setups. Also, the metric for
862 statically configured routes may now be configured. For
863 network interfaces where this is appropriate the peer IP
864 address may now be configured.
866 * systemd-networkd's DHCP client will no longer request
867 broadcasting by default, as this tripped up some networks.
868 For hardware where broadcast is required the feature should
869 be switched back on using RequestBroadcast=yes.
871 * systemd-networkd will now set up IPv4LL addresses (when
872 enabled) even if DHCP is configured successfully.
874 * udev will now default to respect network device names given
875 by the kernel when the kernel indicates that these are
876 predictable. This behavior can be tweaked by changing
877 NamePolicy= in the relevant .link file.
879 * A new library systemd-terminal has been added that
880 implements full TTY stream parsing and rendering. This
881 library is supposed to be used later on for implementing a
882 full userspace VT subsystem, replacing the current kernel
885 * A new tool systemd-journal-upload has been added to push
886 journal data to a remote system running
887 systemd-journal-remote.
889 * journald will no longer forward all local data to another
890 running syslog daemon. This change has been made because
891 rsyslog (which appears to be the most commonly used syslog
892 implementation these days) no longer makes use of this, and
893 instead pulls the data out of the journal on its own. Since
894 forwarding the messages to a non-existent syslog server is
895 more expensive than we assumed we have now turned this
896 off. If you run a syslog server that is not a recent rsyslog
897 version, you have to turn this option on again
898 (ForwardToSyslog= in journald.conf).
900 * journald now optionally supports the LZ4 compressor for
901 larger journal fields. This compressor should perform much
902 better than XZ which was the previous default.
904 * machinectl now shows the IP addresses of local containers,
905 if it knows them, plus the interface name of the container.
907 * A new tool "systemd-escape" has been added that makes it
908 easy to escape strings to build unit names and similar.
910 * sd_notify() messages may now include a new ERRNO= field
911 which is parsed and collected by systemd and shown among the
912 "systemctl status" output for a service.
914 * A new component "systemd-firstboot" has been added that
915 queries the most basic systemd information (timezone,
916 hostname, root password) interactively on first
917 boot. Alternatively it may also be used to provision these
918 things offline on OS images installed into directories.
920 * The default sysctl.d/ snippets will now set
922 net.ipv4.conf.default.promote_secondaries=1
924 This has the benefit of no flushing secondary IP addresses
925 when primary addresses are removed.
927 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
928 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
929 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
930 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
931 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
932 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
933 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
934 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
935 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
936 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
937 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
938 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
939 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
940 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
941 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
943 -- Berlin, 2014-08-19
947 * A new tool systemd-sysusers has been added. This tool
948 creates system users and groups in /etc/passwd and
949 /etc/group, based on static declarative system user/group
950 definitions in /usr/lib/sysusers.d/. This is useful to
951 enable factory resets and volatile systems that boot up with
952 an empty /etc directory, and thus need system users and
953 groups created during early boot. systemd now also ships
954 with two default sysusers.d/ files for the most basic
955 users and groups systemd and the core operating system
958 * A new tmpfiles snippet has been added that rebuilds the
959 essential files in /etc on boot, should they be missing.
961 * A directive for ensuring automatic clean-up of
962 /var/cache/man/ has been removed from the default
963 configuration. This line should now be shipped by the man
964 implementation. The necessary change has been made to the
965 man-db implementation. Note that you need to update your man
966 implementation to one that ships this line, otherwise no
967 automatic clean-up of /var/cache/man will take place.
969 * A new condition ConditionNeedsUpdate= has been added that
970 may conditionalize services to only run when /etc or /var
971 are "older" than the vendor operating system resources in
972 /usr. This is useful for reconstructing or updating /etc
973 after an offline update of /usr or a factory reset, on the
974 next reboot. Services that want to run once after such an
975 update or reset should use this condition and order
976 themselves before the new systemd-update-done.service, which
977 will mark the two directories as fully updated. A number of
978 service files have been added making use of this, to rebuild
979 the udev hardware database, the journald message catalog and
980 dynamic loader cache (ldconfig). The systemd-sysusers tool
981 described above also makes use of this now. With this in
982 place it is now possible to start up a minimal operating
983 system with /etc empty cleanly. For more information on the
984 concepts involved see this recent blog story:
986 http://0pointer.de/blog/projects/stateless.html
988 * A new system group "input" has been introduced, and all
989 input device nodes get this group assigned. This is useful
990 for system-level software to get access to input devices. It
991 complements what is already done for "audio" and "video".
993 * systemd-networkd learnt minimal DHCPv4 server support in
994 addition to the existing DHCPv4 client support. It also
995 learnt DHCPv6 client and IPv6 Router Solicitation client
996 support. The DHCPv4 client gained support for static routes
997 passed in from the server. Note that the [DHCPv4] section
998 known in older systemd-networkd versions has been renamed to
999 [DHCP] and is now also used by the DHCPv6 client. Existing
1000 .network files using settings of this section should be
1001 updated, though compatibility is maintained. Optionally, the
1002 client hostname may now be sent to the DHCP server.
1004 * networkd gained support for vxlan virtual networks as well
1005 as tun/tap and dummy devices.
1007 * networkd gained support for automatic allocation of address
1008 ranges for interfaces from a system-wide pool of
1009 addresses. This is useful for dynamically managing a large
1010 number of interfaces with a single network configuration
1011 file. In particular this is useful to easily assign
1012 appropriate IP addresses to the veth links of a large number
1013 of nspawn instances.
1015 * RPM macros for processing sysusers, sysctl and binfmt
1016 drop-in snippets at package installation time have been
1019 * The /etc/os-release file should now be placed in
1020 /usr/lib/os-release. The old location is automatically
1021 created as symlink. /usr/lib is the more appropriate
1022 location of this file, since it shall actually describe the
1023 vendor operating system shipped in /usr, and not the
1024 configuration stored in /etc.
1026 * .mount units gained a new boolean SloppyOptions= setting
1027 that maps to mount(8)'s -s option which enables permissive
1028 parsing of unknown mount options.
1030 * tmpfiles learnt a new "L+" directive which creates a symlink
1031 but (unlike "L") deletes a pre-existing file first, should
1032 it already exist and not already be the correct
1033 symlink. Similar, "b+", "c+" and "p+" directives have been
1034 added as well, which create block and character devices, as
1035 well as fifos in the filesystem, possibly removing any
1036 pre-existing files of different types.
1038 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
1039 'argument' field (which so far specified the source to
1040 symlink/copy the files from) is now optional. If omitted the
1041 same file os copied from /usr/share/factory/ suffixed by the
1042 full destination path. This is useful for populating /etc
1043 with essential files, by copying them from vendor defaults
1044 shipped in /usr/share/factory/etc.
1046 * A new command "systemctl preset-all" has been added that
1047 applies the service preset settings to all installed unit
1048 files. A new switch --preset-mode= has been added that
1049 controls whether only enable or only disable operations
1052 * A new command "systemctl is-system-running" has been added
1053 that allows checking the overall state of the system, for
1054 example whether it is fully up and running.
1056 * When the system boots up with an empty /etc, the equivalent
1057 to "systemctl preset-all" is executed during early boot, to
1058 make sure all default services are enabled after a factory
1061 * systemd now contains a minimal preset file that enables the
1062 most basic services systemd ships by default.
1064 * Unit files' [Install] section gained a new DefaultInstance=
1065 field for defining the default instance to create if a
1066 template unit is enabled with no instance specified.
1068 * A new passive target cryptsetup-pre.target has been added
1069 that may be used by services that need to make they run and
1070 finish before the first LUKS cryptographic device is set up.
1072 * The /dev/loop-control and /dev/btrfs-control device nodes
1073 are now owned by the "disk" group by default, opening up
1074 access to this group.
1076 * systemd-coredump will now automatically generate a
1077 stack trace of all core dumps taking place on the system,
1078 based on elfutils' libdw library. This stack trace is logged
1081 * systemd-coredump may now optionally store coredumps directly
1082 on disk (in /var/lib/systemd/coredump, possibly compressed),
1083 instead of storing them unconditionally in the journal. This
1084 mode is the new default. A new configuration file
1085 /etc/systemd/coredump.conf has been added to configure this
1086 and other parameters of systemd-coredump.
1088 * coredumpctl gained a new "info" verb to show details about a
1089 specific coredump. A new switch "-1" has also been added
1090 that makes sure to only show information about the most
1091 recent entry instead of all entries. Also, as the tool is
1092 generally useful now the "systemd-" prefix of the binary
1093 name has been removed. Distributions that want to maintain
1094 compatibility with the old name should add a symlink from
1095 the old name to the new name.
1097 * journald's SplitMode= now defaults to "uid". This makes sure
1098 that unprivileged users can access their own coredumps with
1099 coredumpctl without restrictions.
1101 * New kernel command line options "systemd.wants=" (for
1102 pulling an additional unit during boot), "systemd.mask="
1103 (for masking a specific unit for the boot), and
1104 "systemd.debug-shell" (for enabling the debug shell on tty9)
1105 have been added. This is implemented in the new generator
1106 "systemd-debug-generator".
1108 * systemd-nspawn will now by default filter a couple of
1109 syscalls for containers, among them those required for
1110 kernel module loading, direct x86 IO port access, swap
1111 management, and kexec. Most importantly though
1112 open_by_handle_at() is now prohibited for containers,
1113 closing a hole similar to a recently discussed vulnerability
1114 in docker regarding access to files on file hierarchies the
1115 container should normally not have access to. Note that for
1116 nspawn we generally make no security claims anyway (and
1117 this is explicitly documented in the man page), so this is
1118 just a fix for one of the most obvious problems.
1120 * A new man page file-hierarchy(7) has been added that
1121 contains a minimized, modernized version of the file system
1122 layout systemd expects, similar in style to the FHS
1123 specification or hier(5). A new tool systemd-path(1) has
1124 been added to query many of these paths for the local
1127 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
1128 longer done. Since the directory now has a per-user size
1129 limit, and is cleaned on logout this appears unnecessary,
1130 in particular since this now brings the lifecycle of this
1131 directory closer in line with how IPC objects are handled.
1133 * systemd.pc now exports a number of additional directories,
1134 including $libdir (which is useful to identify the library
1135 path for the primary architecture of the system), and a
1136 couple of drop-in directories.
1138 * udev's predictable network interface names now use the dev_port
1139 sysfs attribute, introduced in linux 3.15 instead of dev_id to
1140 distinguish between ports of the same PCI function. dev_id should
1141 only be used for ports using the same HW address, hence the need
1144 * machined has been updated to export the OS version of a
1145 container (read from /etc/os-release and
1146 /usr/lib/os-release) on the bus. This is now shown in
1147 "machinectl status" for a machine.
1149 * A new service setting RestartForceExitStatus= has been
1150 added. If configured to a set of exit signals or process
1151 return values, the service will be restarted when the main
1152 daemon process exits with any of them, regardless of the
1155 * systemctl's -H switch for connecting to remote systemd
1156 machines has been extended so that it may be used to
1157 directly connect to a specific container on the
1158 host. "systemctl -H root@foobar:waldi" will now connect as
1159 user "root" to host "foobar", and then proceed directly to
1160 the container named "waldi". Note that currently you have to
1161 authenticate as user "root" for this to work, as entering
1162 containers is a privileged operation.
1164 Contributions from: Andreas Henriksson, Benjamin Steinwender,
1165 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
1166 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
1167 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
1168 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
1169 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
1170 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
1171 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
1172 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
1173 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
1174 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
1175 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
1177 -- Berlin, 2014-07-03
1181 * As an experimental feature, udev now tries to lock the
1182 disk device node (flock(LOCK_SH|LOCK_NB)) while it
1183 executes events for the disk or any of its partitions.
1184 Applications like partitioning programs can lock the
1185 disk device node (flock(LOCK_EX)) and claim temporary
1186 device ownership that way; udev will entirely skip all event
1187 handling for this disk and its partitions. If the disk
1188 was opened for writing, the close will trigger a partition
1189 table rescan in udev's "watch" facility, and if needed
1190 synthesize "change" events for the disk and all its partitions.
1191 This is now unconditionally enabled, and if it turns out to
1192 cause major problems, we might turn it on only for specific
1193 devices, or might need to disable it entirely. Device Mapper
1194 devices are excluded from this logic.
1196 * We temporarily dropped the "-l" switch for fsck invocations,
1197 since they collide with the flock() logic above. util-linux
1198 upstream has been changed already to avoid this conflict,
1199 and we will readd "-l" as soon as util-linux with this
1200 change has been released.
1202 * The dependency on libattr has been removed. Since a long
1203 time, the extended attribute calls have moved to glibc, and
1204 libattr is thus unnecessary.
1206 * Virtualization detection works without priviliges now. This
1207 means the systemd-detect-virt binary no longer requires
1208 CAP_SYS_PTRACE file capabilities, and our daemons can run
1209 with fewer privileges.
1211 * systemd-networkd now runs under its own "systemd-network"
1212 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
1213 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
1214 loses the ability to write to files owned by root this way.
1216 * Similar, systemd-resolved now runs under its own
1217 "systemd-resolve" user with no capabilities remaining.
1219 * Similar, systemd-bus-proxyd now runs under its own
1220 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
1222 * systemd-networkd gained support for setting up "veth"
1223 virtual ethernet devices for container connectivity, as well
1224 as GRE and VTI tunnels.
1226 * systemd-networkd will no longer automatically attempt to
1227 manually load kernel modules necessary for certain tunnel
1228 transports. Instead, it is assumed the kernel loads them
1229 automatically when required. This only works correctly on
1230 very new kernels. On older kernels, please consider adding
1231 the kernel modules to /etc/modules-load.d/ as a work-around.
1233 * The resolv.conf file systemd-resolved generates has been
1234 moved to /run/systemd/resolve/. If you have a symlink from
1235 /etc/resolv.conf, it might be necessary to correct it.
1237 * Two new service settings, ProtectHome= and ProtectSystem=,
1238 have been added. When enabled, they will make the user data
1239 (such as /home) inaccessible or read-only and the system
1240 (such as /usr) read-only, for specific services. This allows
1241 very light-weight per-service sandboxing to avoid
1242 modifications of user data or system files from
1243 services. These two new switches have been enabled for all
1244 of systemd's long-running services, where appropriate.
1246 * Socket units gained new SocketUser= and SocketGroup=
1247 settings to set the owner user and group of AF_UNIX sockets
1248 and FIFOs in the file system.
1250 * Socket units gained a new RemoveOnStop= setting. If enabled,
1251 all FIFOS and sockets in the file system will be removed
1252 when the specific socket unit is stopped.
1254 * Socket units gained a new Symlinks= setting. It takes a list
1255 of symlinks to create to file system sockets or FIFOs
1256 created by the specific Unix sockets. This is useful to
1257 manage symlinks to socket nodes with the same life-cycle as
1260 * The /dev/log socket and /dev/initctl FIFO have been moved to
1261 /run, and have been replaced by symlinks. This allows
1262 connecting to these facilities even if PrivateDevices=yes is
1263 used for a service (which makes /dev/log itself unavailable,
1264 but /run is left). This also has the benefit of ensuring
1265 that /dev only contains device nodes, directories and
1266 symlinks, and nothing else.
1268 * sd-daemon gained two new calls sd_pid_notify() and
1269 sd_pid_notifyf(). They are similar to sd_notify() and
1270 sd_notifyf(), but allow overriding of the source PID of
1271 notification messages if permissions permit this. This is
1272 useful to send notify messages on behalf of a different
1273 process (for example, the parent process). The
1274 systemd-notify tool has been updated to make use of this
1275 when sending messages (so that notification messages now
1276 originate from the shell script invoking systemd-notify and
1277 not the systemd-notify process itself. This should minimize
1278 a race where systemd fails to associate notification
1279 messages to services when the originating process already
1282 * A new "on-abnormal" setting for Restart= has been added. If
1283 set, it will result in automatic restarts on all "abnormal"
1284 reasons for a process to exit, which includes unclean
1285 signals, core dumps, timeouts and watchdog timeouts, but
1286 does not include clean and unclean exit codes or clean
1287 signals. Restart=on-abnormal is an alternative for
1288 Restart=on-failure for services that shall be able to
1289 terminate and avoid restarts on certain errors, by
1290 indicating so with an unclean exit code. Restart=on-failure
1291 or Restart=on-abnormal is now the recommended setting for
1292 all long-running services.
1294 * If the InaccessibleDirectories= service setting points to a
1295 mount point (or if there are any submounts contained within
1296 it), it is now attempted to completely unmount it, to make
1297 the file systems truly unavailable for the respective
1300 * The ReadOnlyDirectories= service setting and
1301 systemd-nspawn's --read-only parameter are now recursively
1302 applied to all submounts, too.
1304 * Mount units may now be created transiently via the bus APIs.
1306 * The support for SysV and LSB init scripts has been removed
1307 from the systemd daemon itself. Instead, it is now
1308 implemented as a generator that creates native systemd units
1309 from these scripts when needed. This enables us to remove a
1310 substantial amount of legacy code from PID 1, following the
1311 fact that many distributions only ship a very small number
1312 of LSB/SysV init scripts nowadays.
1314 * Privileged Xen (dom0) domains are not considered
1315 virtualization anymore by the virtualization detection
1316 logic. After all, they generally have unrestricted access to
1317 the hardware and usually are used to manage the unprivileged
1320 * systemd-tmpfiles gained a new "C" line type, for copying
1321 files or entire directories.
1323 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
1324 lines. So far, they have been non-globbing versions of the
1325 latter, and have thus been redundant. In future, it is
1326 recommended to only use "z". "m" has hence been removed
1327 from the documentation, even though it stays supported.
1329 * A tmpfiles snippet to recreate the most basic structure in
1330 /var has been added. This is enough to create the /var/run →
1331 /run symlink and create a couple of structural
1332 directories. This allows systems to boot up with an empty or
1333 volatile /var. Of course, while with this change, the core OS
1334 now is capable with dealing with a volatile /var, not all
1335 user services are ready for it. However, we hope that sooner
1336 or later, many service daemons will be changed upstream so
1337 that they are able to automatically create their necessary
1338 directories in /var at boot, should they be missing. This is
1339 the first step to allow state-less systems that only require
1340 the vendor image for /usr to boot.
1342 * systemd-nspawn has gained a new --tmpfs= switch to mount an
1343 empty tmpfs instance to a specific directory. This is
1344 particularly useful for making use of the automatic
1345 reconstruction of /var (see above), by passing --tmpfs=/var.
1347 * Access modes specified in tmpfiles snippets may now be
1348 prefixed with "~", which indicates that they shall be masked
1349 by whether the existing file or directory is currently
1350 writable, readable or executable at all. Also, if specified,
1351 the sgid/suid/sticky bits will be masked for all
1354 * A new passive target unit "network-pre.target" has been
1355 added which is useful for services that shall run before any
1356 network is configured, for example firewall scripts.
1358 * The "floppy" group that previously owned the /dev/fd*
1359 devices is no longer used. The "disk" group is now used
1360 instead. Distributions should probably deprecate usage of
1363 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
1364 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
1365 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
1366 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
1367 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
1368 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
1371 -- Berlin, 2014-06-11
1375 * A new "systemd-timesyncd" daemon has been added for
1376 synchronizing the system clock across the network. It
1377 implements an SNTP client. In contrast to NTP
1378 implementations such as chrony or the NTP reference server,
1379 this only implements a client side, and does not bother with
1380 the full NTP complexity, focusing only on querying time from
1381 one remote server and synchronizing the local clock to
1382 it. Unless you intend to serve NTP to networked clients or
1383 want to connect to local hardware clocks, this simple NTP
1384 client should be more than appropriate for most
1385 installations. The daemon runs with minimal privileges, and
1386 has been hooked up with networkd to only operate when
1387 network connectivity is available. The daemon saves the
1388 current clock to disk every time a new NTP sync has been
1389 acquired, and uses this to possibly correct the system clock
1390 early at bootup, in order to accommodate for systems that
1391 lack an RTC such as the Raspberry Pi and embedded devices,
1392 and to make sure that time monotonically progresses on these
1393 systems, even if it is not always correct. To make use of
1394 this daemon, a new system user and group "systemd-timesync"
1395 needs to be created on installation of systemd.
1397 * The queue "seqnum" interface of libudev has been disabled, as
1398 it was generally incompatible with device namespacing as
1399 sequence numbers of devices go "missing" if the devices are
1400 part of a different namespace.
1402 * "systemctl list-timers" and "systemctl list-sockets" gained
1403 a --recursive switch for showing units of these types also
1404 for all local containers, similar in style to the already
1405 supported --recursive switch for "systemctl list-units".
1407 * A new RebootArgument= setting has been added for service
1408 units, which may be used to specify a kernel reboot argument
1409 to use when triggering reboots with StartLimitAction=.
1411 * A new FailureAction= setting has been added for service
1412 units which may be used to specify an operation to trigger
1413 when a service fails. This works similarly to
1414 StartLimitAction=, but unlike it, controls what is done
1415 immediately rather than only after several attempts to
1416 restart the service in question.
1418 * hostnamed got updated to also expose the kernel name,
1419 release, and version on the bus. This is useful for
1420 executing commands like hostnamectl with the -H switch.
1421 systemd-analyze makes use of this to properly display
1422 details when running non-locally.
1424 * The bootchart tool can now show cgroup information in the
1425 graphs it generates.
1427 * The CFS CPU quota cgroup attribute is now exposed for
1428 services. The new CPUQuota= switch has been added for this
1429 which takes a percentage value. Setting this will have the
1430 result that a service may never get more CPU time than the
1431 specified percentage, even if the machine is otherwise idle.
1433 * systemd-networkd learned IPIP and SIT tunnel support.
1435 * LSB init scripts exposing a dependency on $network will now
1436 get a dependency on network-online.target rather than simply
1437 network.target. This should bring LSB handling closer to
1438 what it was on SysV systems.
1440 * A new fsck.repair= kernel option has been added to control
1441 how fsck shall deal with unclean file systems at boot.
1443 * The (.ini) configuration file parser will now silently
1444 ignore sections whose name begins with "X-". This may be
1445 used to maintain application-specific extension sections in unit
1448 * machined gained a new API to query the IP addresses of
1449 registered containers. "machinectl status" has been updated
1450 to show these addresses in its output.
1452 * A new call sd_uid_get_display() has been added to the
1453 sd-login APIs for querying the "primary" session of a
1454 user. The "primary" session of the user is elected from the
1455 user's sessions and generally a graphical session is
1456 preferred over a text one.
1458 * A minimal systemd-resolved daemon has been added. It
1459 currently simply acts as a companion to systemd-networkd and
1460 manages resolv.conf based on per-interface DNS
1461 configuration, possibly supplied via DHCP. In the long run
1462 we hope to extend this into a local DNSSEC enabled DNS and
1465 * The systemd-networkd-wait-online tool is now enabled by
1466 default. It will delay network-online.target until a network
1467 connection has been configured. The tool primarily integrates
1468 with networkd, but will also make a best effort to make sense
1469 of network configuration performed in some other way.
1471 * Two new service options StartupCPUShares= and
1472 StartupBlockIOWeight= have been added that work similarly to
1473 CPUShares= and BlockIOWeight= however only apply during
1474 system startup. This is useful to prioritize certain services
1475 differently during bootup than during normal runtime.
1477 * hostnamed has been changed to prefer the statically
1478 configured hostname in /etc/hostname (unless set to
1479 'localhost' or empty) over any dynamic one supplied by
1480 dhcp. With this change, the rules for picking the hostname
1481 match more closely the rules of other configuration settings
1482 where the local administrator's configuration in /etc always
1483 overrides any other settings.
1485 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
1486 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
1487 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
1488 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
1489 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
1490 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
1491 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
1492 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
1493 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
1494 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
1495 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
1496 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
1497 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
1498 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
1499 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
1500 Lindskog, WaLyong Cho, Will Woods, Zbigniew
1503 -- Beijing, 2014-05-28
1507 * When restoring the screen brightness at boot, stay away from
1508 the darkest setting or from the lowest 5% of the available
1509 range, depending on which is the larger value of both. This
1510 should effectively protect the user from rebooting into a
1511 black screen, should the brightness have been set to minimum
1514 * sd-login gained a new sd_machine_get_class() call to
1515 determine the class ("vm" or "container") of a machine
1516 registered with machined.
1518 * sd-login gained new calls
1519 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
1520 to query the identity of the peer of a local AF_UNIX
1521 connection. They operate similarly to their sd_pid_get_xyz()
1524 * PID 1 will now maintain a system-wide system state engine
1525 with the states "starting", "running", "degraded",
1526 "maintenance", "stopping". These states are bound to system
1527 startup, normal runtime, runtime with at least one failed
1528 service, rescue/emergency mode and system shutdown. This
1529 state is shown in the "systemctl status" output when no unit
1530 name is passed. It is useful to determine system state, in
1531 particularly when doing so for many systems or containers at
1534 * A new command "list-machines" has been added to "systemctl"
1535 that lists all local OS containers and shows their system
1536 state (see above), if systemd runs inside of them.
1538 * systemctl gained a new "-r" switch to recursively enumerate
1539 units on all local containers, when used with the
1540 "list-unit" command (which is the default one that is
1541 executed when no parameters are specified).
1543 * The GPT automatic partition discovery logic will now honour
1544 two GPT partition flags: one may be set on a partition to
1545 cause it to be mounted read-only, and the other may be set
1546 on a partition to ignore it during automatic discovery.
1548 * Two new GPT type UUIDs have been added for automatic root
1549 partition discovery, for 32-bit and 64-bit ARM. This is not
1550 particularly useful for discovering the root directory on
1551 these architectures during bare-metal boots (since UEFI is
1552 not common there), but still very useful to allow booting of
1553 ARM disk images in nspawn with the -i option.
1555 * MAC addresses of interfaces created with nspawn's
1556 --network-interface= switch will now be generated from the
1557 machine name, and thus be stable between multiple invocations
1560 * logind will now automatically remove all IPC objects owned
1561 by a user if she or he fully logs out. This makes sure that
1562 users who are logged out cannot continue to consume IPC
1563 resources. This covers SysV memory, semaphores and message
1564 queues as well as POSIX shared memory and message
1565 queues. Traditionally, SysV and POSIX IPC had no life-cycle
1566 limits. With this functionality, that is corrected. This may
1567 be turned off by using the RemoveIPC= switch of logind.conf.
1569 * The systemd-machine-id-setup and tmpfiles tools gained a
1570 --root= switch to operate on a specific root directory,
1573 * journald can now forward logged messages to the TTYs of all
1574 logged in users ("wall"). This is the default for all
1575 emergency messages now.
1577 * A new tool systemd-journal-remote has been added to stream
1578 journal log messages across the network.
1580 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
1581 controller trees are mounted into it. Note that the
1582 directories mounted beneath it are not read-only. This is a
1583 security measure and is particularly useful because glibc
1584 actually includes a search logic to pick any tmpfs it can
1585 find to implement shm_open() if /dev/shm is not available
1586 (which it might very well be in namespaced setups).
1588 * machinectl gained a new "poweroff" command to cleanly power
1589 down a local OS container.
1591 * The PrivateDevices= unit file setting will now also drop the
1592 CAP_MKNOD capability from the capability bound set, and
1593 imply DevicePolicy=closed.
1595 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
1596 comprehensively on all long-running systemd services where
1597 this is appropriate.
1599 * systemd-udevd will now run in a disassociated mount
1600 namespace. To mount directories from udev rules, make sure to
1601 pull in mount units via SYSTEMD_WANTS properties.
1603 * The kdbus support gained support for uploading policy into
1604 the kernel. sd-bus gained support for creating "monitoring"
1605 connections that can eavesdrop into all bus communication
1606 for debugging purposes.
1608 * Timestamps may now be specified in seconds since the UNIX
1609 epoch Jan 1st, 1970 by specifying "@" followed by the value
1612 * Native tcpwrap support in systemd has been removed. tcpwrap
1613 is old code, not really maintained anymore and has serious
1614 shortcomings, and better options such as firewalls
1615 exist. For setups that require tcpwrap usage, please
1616 consider invoking your socket-activated service via tcpd,
1617 like on traditional inetd.
1619 * A new system.conf configuration option
1620 DefaultTimerAccuracySec= has been added that controls the
1621 default AccuracySec= setting of .timer units.
1623 * Timer units gained a new WakeSystem= switch. If enabled,
1624 timers configured this way will cause the system to resume
1625 from system suspend (if the system supports that, which most
1628 * Timer units gained a new Persistent= switch. If enabled,
1629 timers configured this way will save to disk when they have
1630 been last triggered. This information is then used on next
1631 reboot to possible execute overdue timer events, that
1632 could not take place because the system was powered off.
1633 This enables simple anacron-like behaviour for timer units.
1635 * systemctl's "list-timers" will now also list the time a
1636 timer unit was last triggered in addition to the next time
1637 it will be triggered.
1639 * systemd-networkd will now assign predictable IPv4LL
1640 addresses to its local interfaces.
1642 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
1643 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
1644 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
1645 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
1646 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
1647 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
1648 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
1649 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
1652 -- Berlin, 2014-03-25
1656 * A new unit file setting RestrictAddressFamilies= has been
1657 added to restrict which socket address families unit
1658 processes gain access to. This takes address family names
1659 like "AF_INET" or "AF_UNIX", and is useful to minimize the
1660 attack surface of services via exotic protocol stacks. This
1661 is built on seccomp system call filters.
1663 * Two new unit file settings RuntimeDirectory= and
1664 RuntimeDirectoryMode= have been added that may be used to
1665 manage a per-daemon runtime directories below /run. This is
1666 an alternative for setting up directory permissions with
1667 tmpfiles snippets, and has the advantage that the runtime
1668 directory's lifetime is bound to the daemon runtime and that
1669 the daemon starts up with an empty directory each time. This
1670 is particularly useful when writing services that drop
1671 privileges using the User= or Group= setting.
1673 * The DeviceAllow= unit setting now supports globbing for
1674 matching against device group names.
1676 * The systemd configuration file system.conf gained new
1677 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
1678 DefaultMemoryAccounting= to globally turn on/off accounting
1679 for specific resources (cgroups) for all units. These
1680 settings may still be overridden individually in each unit
1683 * systemd-gpt-auto-generator is now able to discover /srv and
1684 root partitions in addition to /home and swap partitions. It
1685 also supports LUKS-encrypted partitions now. With this in
1686 place, automatic discovery of partitions to mount following
1687 the Discoverable Partitions Specification
1688 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
1689 is now a lot more complete. This allows booting without
1690 /etc/fstab and without root= on the kernel command line on
1691 systems prepared appropriately.
1693 * systemd-nspawn gained a new --image= switch which allows
1694 booting up disk images and Linux installations on any block
1695 device that follow the Discoverable Partitions Specification
1696 (see above). This means that installations made with
1697 appropriately updated installers may now be started and
1698 deployed using container managers, completely
1699 unmodified. (We hope that libvirt-lxc will add support for
1700 this feature soon, too.)
1702 * systemd-nspawn gained a new --network-macvlan= setting to
1703 set up a private macvlan interface for the
1704 container. Similarly, systemd-networkd gained a new
1705 Kind=macvlan setting in .netdev files.
1707 * systemd-networkd now supports configuring local addresses
1710 * A new tool systemd-network-wait-online has been added to
1711 synchronously wait for network connectivity using
1714 * The sd-bus.h bus API gained a new sd_bus_track object for
1715 tracking the life-cycle of bus peers. Note that sd-bus.h is
1716 still not a public API though (unless you specify
1717 --enable-kdbus on the configure command line, which however
1718 voids your warranty and you get no API stability guarantee).
1720 * The $XDG_RUNTIME_DIR runtime directories for each user are
1721 now individual tmpfs instances, which has the benefit of
1722 introducing separate pools for each user, with individual
1723 size limits, and thus making sure that unprivileged clients
1724 can no longer negatively impact the system or other users by
1725 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
1726 RuntimeDirectorySize= has been introduced that allows
1727 controlling the default size limit for all users. It
1728 defaults to 10% of the available physical memory. This is no
1729 replacement for quotas on tmpfs though (which the kernel
1730 still does not support), as /dev/shm and /tmp are still
1731 shared resources used by both the system and unprivileged
1734 * logind will now automatically turn off automatic suspending
1735 on laptop lid close when more than one display is
1736 connected. This was previously expected to be implemented
1737 individually in desktop environments (such as GNOME),
1738 however has been added to logind now, in order to fix a
1739 boot-time race where a desktop environment might not have
1740 been started yet and thus not been able to take an inhibitor
1741 lock at the time where logind already suspends the system
1742 due to a closed lid.
1744 * logind will now wait at least 30s after each system
1745 suspend/resume cycle, and 3min after system boot before
1746 suspending the system due to a closed laptop lid. This
1747 should give USB docking stations and similar enough time to
1748 be probed and configured after system resume and boot in
1749 order to then act as suspend blocker.
1751 * systemd-run gained a new --property= setting which allows
1752 initialization of resource control properties (and others)
1753 for the created scope or service unit. Example: "systemd-run
1754 --property=BlockIOWeight=10 updatedb" may be used to run
1755 updatedb at a low block IO scheduling weight.
1757 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
1758 now also work in --scope mode.
1760 * When systemd is compiled with kdbus support, basic support
1761 for enforced policies is now in place. (Note that enabling
1762 kdbus still voids your warranty and no API compatibility
1765 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
1766 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1767 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
1768 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
1769 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
1770 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
1771 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
1772 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
1773 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
1774 Zbigniew Jędrzejewski-Szmek
1776 -- Berlin, 2014-03-12
1780 * systemd will now relabel /dev after loading the SMACK policy
1781 according to SMACK rules.
1783 * A new unit file option AppArmorProfile= has been added to
1784 set the AppArmor profile for the processes of a unit.
1786 * A new condition check ConditionArchitecture= has been added
1787 to conditionalize units based on the system architecture, as
1788 reported by uname()'s "machine" field.
1790 * systemd-networkd now supports matching on the system
1791 virtualization, architecture, kernel command line, host name
1794 * logind is now a lot more aggressive when suspending the
1795 machine due to a closed laptop lid. Instead of acting only
1796 on the lid close action, it will continuously watch the lid
1797 status and act on it. This is useful for laptops where the
1798 power button is on the outside of the chassis so that it can
1799 be reached without opening the lid (such as the Lenovo
1800 Yoga). On those machines, logind will now immediately
1801 re-suspend the machine if the power button has been
1802 accidentally pressed while the laptop was suspended and in a
1803 backpack or similar.
1805 * logind will now watch SW_DOCK switches and inhibit reaction
1806 to the lid switch if it is pressed. This means that logind
1807 will not suspend the machine anymore if the lid is closed
1808 and the system is docked, if the laptop supports SW_DOCK
1809 notifications via the input layer. Note that ACPI docking
1810 stations do not generate this currently. Also note that this
1811 logic is usually not fully sufficient and Desktop
1812 Environments should take a lid switch inhibitor lock when an
1813 external display is connected, as systemd will not watch
1816 * nspawn will now make use of the devices cgroup controller by
1817 default, and only permit creation of and access to the usual
1818 API device nodes like /dev/null or /dev/random, as well as
1819 access to (but not creation of) the pty devices.
1821 * We will now ship a default .network file for
1822 systemd-networkd that automatically configures DHCP for
1823 network interfaces created by nspawn's --network-veth or
1824 --network-bridge= switches.
1826 * systemd will now understand the usual M, K, G, T suffixes
1827 according to SI conventions (i.e. to the base 1000) when
1828 referring to throughput and hardware metrics. It will stay
1829 with IEC conventions (i.e. to the base 1024) for software
1830 metrics, according to what is customary according to
1831 Wikipedia. We explicitly document which base applies for
1832 each configuration option.
1834 * The DeviceAllow= setting in unit files now supports a syntax
1835 to whitelist an entire group of devices node majors at once,
1836 based on the /proc/devices listing. For example, with the
1837 string "char-pts", it is now possible to whitelist all
1838 current and future pseudo-TTYs at once.
1840 * sd-event learned a new "post" event source. Event sources of
1841 this type are triggered by the dispatching of any event
1842 source of a type that is not "post". This is useful for
1843 implementing clean-up and check event sources that are
1844 triggered by other work being done in the program.
1846 * systemd-networkd is no longer statically enabled, but uses
1847 the usual [Install] sections so that it can be
1848 enabled/disabled using systemctl. It still is enabled by
1851 * When creating a veth interface pair with systemd-nspawn, the
1852 host side will now be prefixed with "vb-" if
1853 --network-bridge= is used, and with "ve-" if --network-veth
1854 is used. This way, it is easy to distinguish these cases on
1855 the host, for example to apply different configuration to
1856 them with systemd-networkd.
1858 * The compatibility libraries for libsystemd-journal.so,
1859 libsystem-id128.so, libsystemd-login.so and
1860 libsystemd-daemon.so do not make use of IFUNC
1861 anymore. Instead, we now build libsystemd.so multiple times
1862 under these alternative names. This means that the footprint
1863 is drastically increased, but given that these are
1864 transitional compatibility libraries, this should not matter
1865 much. This change has been made necessary to support the ARM
1866 platform for these compatibility libraries, as the ARM
1867 toolchain is not really at the same level as the toolchain
1868 for other architectures like x86 and does not support
1869 IFUNC. Please make sure to use --enable-compat-libs only
1870 during a transitional period!
1872 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
1873 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1874 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
1875 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
1876 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
1877 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
1878 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
1879 Zbigniew Jędrzejewski-Szmek
1881 -- Berlin, 2014-02-24
1885 * A new component "systemd-networkd" has been added that can
1886 be used to configure local network interfaces statically or
1887 via DHCP. It is capable of bringing up bridges, VLANs, and
1888 bonding. Currently, no hook-ups for interactive network
1889 configuration are provided. Use this for your initrd,
1890 container, embedded, or server setup if you need a simple,
1891 yet powerful, network configuration solution. This
1892 configuration subsystem is quite nifty, as it allows wildcard
1893 hotplug matching in interfaces. For example, with a single
1894 configuration snippet, you can configure that all Ethernet
1895 interfaces showing up are automatically added to a bridge,
1896 or similar. It supports link-sensing and more.
1898 * A new tool "systemd-socket-proxyd" has been added which can
1899 act as a bidirectional proxy for TCP sockets. This is
1900 useful for adding socket activation support to services that
1901 do not actually support socket activation, including virtual
1902 machines and the like.
1904 * Add a new tool to save/restore rfkill state on
1907 * Save/restore state of keyboard backlights in addition to
1908 display backlights on shutdown/boot.
1910 * udev learned a new SECLABEL{} construct to label device
1911 nodes with a specific security label when they appear. For
1912 now, only SECLABEL{selinux} is supported, but the syntax is
1913 prepared for additional security frameworks.
1915 * udev gained a new scheme to configure link-level attributes
1916 from files in /etc/systemd/network/*.link. These files can
1917 match against MAC address, device path, driver name and type,
1918 and will apply attributes like the naming policy, link speed,
1919 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
1920 address assignment policy (randomized, ...).
1922 * The configuration of network interface naming rules for
1923 "permanent interface names" has changed: a new NamePolicy=
1924 setting in the [Link] section of .link files determines the
1925 priority of possible naming schemes (onboard, slot, mac,
1926 path). The default value of this setting is determined by
1927 /usr/lib/net/links/99-default.link. Old
1928 80-net-name-slot.rules udev configuration file has been
1929 removed, so local configuration overriding this file should
1930 be adapated to override 99-default.link instead.
1932 * When the User= switch is used in a unit file, also
1933 initialize $SHELL= based on the user database entry.
1935 * systemd no longer depends on libdbus. All communication is
1936 now done with sd-bus, systemd's low-level bus library
1939 * kdbus support has been added to PID 1 itself. When kdbus is
1940 enabled, this causes PID 1 to set up the system bus and
1941 enable support for a new ".busname" unit type that
1942 encapsulates bus name activation on kdbus. It works a little
1943 bit like ".socket" units, except for bus names. A new
1944 generator has been added that converts classic dbus1 service
1945 activation files automatically into native systemd .busname
1948 * sd-bus: add a light-weight vtable implementation that allows
1949 defining objects on the bus with a simple static const
1950 vtable array of its methods, signals and properties.
1952 * systemd will not generate or install static dbus
1953 introspection data anymore to /usr/share/dbus-1/interfaces,
1954 as the precise format of these files is unclear, and
1955 nothing makes use of it.
1957 * A proxy daemon is now provided to proxy clients connecting
1958 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
1959 compatibility with classic D-Bus.
1961 * A bus driver implementation has been added that supports the
1962 classic D-Bus bus driver calls on kdbus, also for
1963 compatibility purposes.
1965 * A new API "sd-event.h" has been added that implements a
1966 minimal event loop API built around epoll. It provides a
1967 couple of features that direct epoll usage is lacking:
1968 prioritization of events, scales to large numbers of timer
1969 events, per-event timer slack (accuracy), system-wide
1970 coalescing of timer events, exit handlers, watchdog
1971 supervision support using systemd's sd_notify() API, child
1974 * A new API "sd-rntl.h" has been added that provides an API
1975 around the route netlink interface of the kernel, similar in
1976 style to "sd-bus.h".
1978 * A new API "sd-dhcp-client.h" has been added that provides a
1979 small DHCPv4 client-side implementation. This is used by
1982 * There is a new kernel command line option
1983 "systemd.restore_state=0|1". When set to "0", none of the
1984 systemd tools will restore saved runtime state to hardware
1985 devices. More specifically, the rfkill and backlight states
1988 * The FsckPassNo= compatibility option in mount/service units
1989 has been removed. The fstab generator will now add the
1990 necessary dependencies automatically, and does not require
1991 PID1's support for that anymore.
1993 * journalctl gained a new switch, --list-boots, that lists
1994 recent boots with their times and boot IDs.
1996 * The various tools like systemctl, loginctl, timedatectl,
1997 busctl, systemd-run, ... have gained a new switch "-M" to
1998 connect to a specific, local OS container (as direct
1999 connection, without requiring SSH). This works on any
2000 container that is registered with machined, such as those
2001 created by libvirt-lxc or nspawn.
2003 * systemd-run and systemd-analyze also gained support for "-H"
2004 to connect to remote hosts via SSH. This is particularly
2005 useful for systemd-run because it enables queuing of jobs
2006 onto remote systems.
2008 * machinectl gained a new command "login" to open a getty
2009 login in any local container. This works with any container
2010 that is registered with machined (such as those created by
2011 libvirt-lxc or nspawn), and which runs systemd inside.
2013 * machinectl gained a new "reboot" command that may be used to
2014 trigger a reboot on a specific container that is registered
2015 with machined. This works on any container that runs an init
2016 system of some kind.
2018 * systemctl gained a new "list-timers" command to print a nice
2019 listing of installed timer units with the times they elapse
2022 * Alternative reboot() parameters may now be specified on the
2023 "systemctl reboot" command line and are passed to the
2024 reboot() system call.
2026 * systemctl gained a new --job-mode= switch to configure the
2027 mode to queue a job with. This is a more generic version of
2028 --fail, --irreversible, and --ignore-dependencies, which are
2029 still available but not advertised anymore.
2031 * /etc/systemd/system.conf gained new settings to configure
2032 various default timeouts of units, as well as the default
2033 start limit interval and burst. These may still be overridden
2036 * PID1 will now export on the bus profile data of the security
2037 policy upload process (such as the SELinux policy upload to
2040 * journald: when forwarding logs to the console, include
2041 timestamps (following the setting in
2042 /sys/module/printk/parameters/time).
2044 * OnCalendar= in timer units now understands the special
2045 strings "yearly" and "annually". (Both are equivalent)
2047 * The accuracy of timer units is now configurable with the new
2048 AccuracySec= setting. It defaults to 1min.
2050 * A new dependency type JoinsNamespaceOf= has been added that
2051 allows running two services within the same /tmp and network
2052 namespace, if PrivateNetwork= or PrivateTmp= are used.
2054 * A new command "cat" has been added to systemctl. It outputs
2055 the original unit file of a unit, and concatenates the
2056 contents of additional "drop-in" unit file snippets, so that
2057 the full configuration is shown.
2059 * systemctl now supports globbing on the various "list-xyz"
2060 commands, like "list-units" or "list-sockets", as well as on
2061 those commands which take multiple unit names.
2063 * journalctl's --unit= switch gained support for globbing.
2065 * All systemd daemons now make use of the watchdog logic so
2066 that systemd automatically notices when they hang.
2068 * If the $container_ttys environment variable is set,
2069 getty-generator will automatically spawn a getty for each
2070 listed tty. This is useful for container managers to request
2071 login gettys to be spawned on as many ttys as needed.
2073 * %h, %s, %U specifier support is not available anymore when
2074 used in unit files for PID 1. This is because NSS calls are
2075 not safe from PID 1. They stay available for --user
2076 instances of systemd, and as special case for the root user.
2078 * loginctl gained a new "--no-legend" switch to turn off output
2081 * The "sd-login.h" API gained three new calls:
2082 sd_session_is_remote(), sd_session_get_remote_user(),
2083 sd_session_get_remote_host() to query information about
2086 * The udev hardware database now also carries vendor/product
2087 information of SDIO devices.
2089 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
2090 determine whether watchdog notifications are requested by
2093 * Socket-activated per-connection services now include a
2094 short description of the connection parameters in the
2097 * tmpfiles gained a new "--boot" option. When this is not used,
2098 only lines where the command character is not suffixed with
2099 "!" are executed. When this option is specified, those
2100 options are executed too. This partitions tmpfiles
2101 directives into those that can be safely executed at any
2102 time, and those which should be run only at boot (for
2103 example, a line that creates /run/nologin).
2105 * A new API "sd-resolve.h" has been added which provides a simple
2106 asynchronous wrapper around glibc NSS host name resolution
2107 calls, such as getaddrinfo(). In contrast to glibc's
2108 getaddrinfo_a(), it does not use signals. In contrast to most
2109 other asynchronous name resolution libraries, this one does
2110 not reimplement DNS, but reuses NSS, so that alternate
2111 host name resolution systems continue to work, such as mDNS,
2112 LDAP, etc. This API is based on libasyncns, but it has been
2113 cleaned up for inclusion in systemd.
2115 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
2116 "sd-daemon.h" are no longer found in individual libraries
2117 libsystemd-journal.so, libsystemd-login.so,
2118 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
2119 merged them into a single library, libsystemd.so, which
2120 provides all symbols. The reason for this is cyclic
2121 dependencies, as these libraries tend to use each other's
2122 symbols. So far, we have managed to workaround that by linking
2123 a copy of a good part of our code into each of these
2124 libraries again and again, which, however, makes certain
2125 things hard to do, like sharing static variables. Also, it
2126 substantially increases footprint. With this change, there
2127 is only one library for the basic APIs systemd
2128 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
2129 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
2130 library as well, however are subject to the --enable-kdbus
2131 switch (see below). Note that "sd-dhcp-client.h" is not part
2132 of this library (this is because it only consumes, never
2133 provides, services of/to other APIs). To make the transition
2134 easy from the separate libraries to the unified one, we
2135 provide the --enable-compat-libs compile-time switch which
2136 will generate stub libraries that are compatible with the
2137 old ones but redirect all calls to the new one.
2139 * All of the kdbus logic and the new APIs "sd-bus.h",
2140 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
2141 and "sd-utf8.h" are compile-time optional via the
2142 "--enable-kdbus" switch, and they are not compiled in by
2143 default. To make use of kdbus, you have to explicitly enable
2144 the switch. Note however, that neither the kernel nor the
2145 userspace API for all of this is considered stable yet. We
2146 want to maintain the freedom to still change the APIs for
2147 now. By specifying this build-time switch, you acknowledge
2148 that you are aware of the instability of the current
2151 * Also, note that while kdbus is pretty much complete,
2152 it lacks one thing: proper policy support. This means you
2153 can build a fully working system with all features; however,
2154 it will be highly insecure. Policy support will be added in
2155 one of the next releases, at the same time that we will
2156 declare the APIs stable.
2158 * When the kernel command line argument "kdbus" is specified,
2159 systemd will automatically load the kdbus.ko kernel module. At
2160 this stage of development, it is only useful for testing kdbus
2161 and should not be used in production. Note: if "--enable-kdbus"
2162 is specified, and the kdbus.ko kernel module is available, and
2163 "kdbus" is added to the kernel command line, the entire system
2164 runs with kdbus instead of dbus-daemon, with the above mentioned
2165 problem of missing the system policy enforcement. Also a future
2166 version of kdbus.ko or a newer systemd will not be compatible with
2167 each other, and will unlikely be able to boot the machine if only
2168 one of them is updated.
2170 * systemctl gained a new "import-environment" command which
2171 uploads the caller's environment (or parts thereof) into the
2172 service manager so that it is inherited by services started
2173 by the manager. This is useful to upload variables like
2174 $DISPLAY into the user service manager.
2176 * A new PrivateDevices= switch has been added to service units
2177 which allows running a service with a namespaced /dev
2178 directory that does not contain any device nodes for
2179 physical devices. More specifically, it only includes devices
2180 such as /dev/null, /dev/urandom, and /dev/zero which are API
2183 * logind has been extended to support behaviour like VT
2184 switching on seats that do not support a VT. This makes
2185 multi-session available on seats that are not the first seat
2186 (seat0), and on systems where kernel support for VTs has
2187 been disabled at compile-time.
2189 * If a process holds a delay lock for system sleep or shutdown
2190 and fails to release it in time, we will now log its
2191 identity. This makes it easier to identify processes that
2192 cause slow suspends or power-offs.
2194 * When parsing /etc/crypttab, support for a new key-slot=
2195 option as supported by Debian is added. It allows indicating
2196 which LUKS slot to use on disk, speeding up key loading.
2198 * The sd_journald_sendv() API call has been checked and
2199 officially declared to be async-signal-safe so that it may
2200 be invoked from signal handlers for logging purposes.
2202 * Boot-time status output is now enabled automatically after a
2203 short timeout if boot does not progress, in order to give
2204 the user an indication what she or he is waiting for.
2206 * The boot-time output has been improved to show how much time
2207 remains until jobs expire.
2209 * The KillMode= switch in service units gained a new possible
2210 value "mixed". If set, and the unit is shut down, then the
2211 initial SIGTERM signal is sent only to the main daemon
2212 process, while the following SIGKILL signal is sent to
2213 all remaining processes of the service.
2215 * When a scope unit is registered, a new property "Controller"
2216 may be set. If set to a valid bus name, systemd will send a
2217 RequestStop() signal to this name when it would like to shut
2218 down the scope. This may be used to hook manager logic into
2219 the shutdown logic of scope units. Also, scope units may now
2220 be put in a special "abandoned" state, in which case the
2221 manager process which created them takes no further
2222 responsibilities for it.
2224 * When reading unit files, systemd will now verify
2225 the access mode of these files, and warn about certain
2226 suspicious combinations. This has been added to make it
2227 easier to track down packaging bugs where unit files are
2228 marked executable or world-writable.
2230 * systemd-nspawn gained a new "--setenv=" switch to set
2231 container-wide environment variables. The similar option in
2232 systemd-activate was renamed from "--environment=" to
2233 "--setenv=" for consistency.
2235 * systemd-nspawn has been updated to create a new kdbus domain
2236 for each container that is invoked, thus allowing each
2237 container to have its own set of system and user buses,
2238 independent of the host.
2240 * systemd-nspawn gained a new --drop-capability= switch to run
2241 the container with less capabilities than the default. Both
2242 --drop-capability= and --capability= now take the special
2243 string "all" for dropping or keeping all capabilities.
2245 * systemd-nspawn gained new switches for executing containers
2246 with specific SELinux labels set.
2248 * systemd-nspawn gained a new --quiet switch to not generate
2249 any additional output but the container's own console
2252 * systemd-nspawn gained a new --share-system switch to run a
2253 container without PID namespacing enabled.
2255 * systemd-nspawn gained a new --register= switch to control
2256 whether the container is registered with systemd-machined or
2257 not. This is useful for containers that do not run full
2258 OS images, but only specific apps.
2260 * systemd-nspawn gained a new --keep-unit which may be used
2261 when invoked as the only program from a service unit, and
2262 results in registration of the unit service itself in
2263 systemd-machined, instead of a newly opened scope unit.
2265 * systemd-nspawn gained a new --network-interface= switch for
2266 moving arbitrary interfaces to the container. The new
2267 --network-veth switch creates a virtual Ethernet connection
2268 between host and container. The new --network-bridge=
2269 switch then allows assigning the host side of this virtual
2270 Ethernet connection to a bridge device.
2272 * systemd-nspawn gained a new --personality= switch for
2273 setting the kernel personality for the container. This is
2274 useful when running a 32-bit container on a 64-bit host. A
2275 similar option Personality= is now also available for service
2278 * logind will now also track a "Desktop" identifier for each
2279 session which encodes the desktop environment of it. This is
2280 useful for desktop environments that want to identify
2281 multiple running sessions of itself easily.
2283 * A new SELinuxContext= setting for service units has been
2284 added that allows setting a specific SELinux execution
2285 context for a service.
2287 * Most systemd client tools will now honour $SYSTEMD_LESS for
2288 settings of the "less" pager. By default, these tools will
2289 override $LESS to allow certain operations to work, such as
2290 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
2291 influence this logic.
2293 * systemd's "seccomp" hook-up has been changed to make use of
2294 the libseccomp library instead of using its own
2295 implementation. This has benefits for portability among
2298 * For usage together with SystemCallFilter=, a new
2299 SystemCallErrorNumber= setting has been introduced that
2300 allows configuration of a system error number to be returned
2301 on filtered system calls, instead of immediately killing the
2302 process. Also, SystemCallArchitectures= has been added to
2303 limit access to system calls of a particular architecture
2304 (in order to turn off support for unused secondary
2305 architectures). There is also a global
2306 SystemCallArchitectures= setting in system.conf now to turn
2307 off support for non-native system calls system-wide.
2309 * systemd requires a kernel with a working name_to_handle_at(),
2310 please see the kernel config requirements in the README file.
2312 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
2313 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
2314 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
2315 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
2316 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
2317 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
2318 Elia Pinto, Florian Weimer, George McCollister, Goffredo
2319 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
2320 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
2321 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
2322 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
2323 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
2324 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
2325 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
2326 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
2327 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
2328 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
2329 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
2330 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
2331 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
2332 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
2333 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
2334 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
2335 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
2337 -- Berlin, 2014-02-20
2341 * logind has gained support for facilitating privileged input
2342 and drm device access for unprivileged clients. This work is
2343 useful to allow Wayland display servers (and similar
2344 programs, such as kmscon) to run under the user's ID and
2345 access input and drm devices which are normally
2346 protected. When this is used (and the kernel is new enough)
2347 logind will "mute" IO on the file descriptors passed to
2348 Wayland as long as it is in the background and "unmute" it
2349 if it returns into the foreground. This allows secure
2350 session switching without allowing background sessions to
2351 eavesdrop on input and display data. This also introduces
2352 session switching support if VT support is turned off in the
2353 kernel, and on seats that are not seat0.
2355 * A new kernel command line option luks.options= is understood
2356 now which allows specifying LUKS options for usage for LUKS
2357 encrypted partitions specified with luks.uuid=.
2359 * tmpfiles.d(5) snippets may now use specifier expansion in
2360 path names. More specifically %m, %b, %H, %v, are now
2361 replaced by the local machine id, boot id, hostname, and
2362 kernel version number.
2364 * A new tmpfiles.d(5) command "m" has been introduced which
2365 may be used to change the owner/group/access mode of a file
2366 or directory if it exists, but do nothing if it does not.
2368 * This release removes high-level support for the
2369 MemorySoftLimit= cgroup setting. The underlying kernel
2370 cgroup attribute memory.soft_limit= is currently badly
2371 designed and likely to be removed from the kernel API in its
2372 current form, hence we should not expose it for now.
2374 * The memory.use_hierarchy cgroup attribute is now enabled for
2375 all cgroups systemd creates in the memory cgroup
2376 hierarchy. This option is likely to be come the built-in
2377 default in the kernel anyway, and the non-hierarchical mode
2378 never made much sense in the intrinsically hierarchical
2381 * A new field _SYSTEMD_SLICE= is logged along with all journal
2382 messages containing the slice a message was generated
2383 from. This is useful to allow easy per-customer filtering of
2384 logs among other things.
2386 * systemd-journald will no longer adjust the group of journal
2387 files it creates to the "systemd-journal" group. Instead we
2388 rely on the journal directory to be owned by the
2389 "systemd-journal" group, and its setgid bit set, so that the
2390 kernel file system layer will automatically enforce that
2391 journal files inherit this group assignment. The reason for
2392 this change is that we cannot allow NSS look-ups from
2393 journald which would be necessary to resolve
2394 "systemd-journal" to a numeric GID, because this might
2395 create deadlocks if NSS involves synchronous queries to
2396 other daemons (such as nscd, or sssd) which in turn are
2397 logging clients of journald and might block on it, which
2398 would then dead lock. A tmpfiles.d(5) snippet included in
2399 systemd will make sure the setgid bit and group are
2400 properly set on the journal directory if it exists on every
2401 boot. However, we recommend adjusting it manually after
2402 upgrades too (or from RPM scriptlets), so that the change is
2403 not delayed until next reboot.
2405 * Backlight and random seed files in /var/lib/ have moved into
2406 the /var/lib/systemd/ directory, in order to centralize all
2407 systemd generated files in one directory.
2409 * Boot time performance measurements (as displayed by
2410 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
2411 performance information if that's available to determine how
2412 much time BIOS and boot loader initialization required. With
2413 a sufficiently new BIOS you hence no longer need to boot
2414 with Gummiboot to get access to such information.
2416 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
2417 Cristian Rodríguez, Dave Reisner, David Herrmann, David
2418 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
2419 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
2420 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
2421 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
2422 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2424 -- Berlin, 2013-10-02
2428 * The Restart= option for services now understands a new
2429 on-watchdog setting, which will restart the service
2430 automatically if the service stops sending out watchdog keep
2431 alive messages (as configured with WatchdogSec=).
2433 * The getty generator (which is responsible for bringing up a
2434 getty on configured serial consoles) will no longer only
2435 start a getty on the primary kernel console but on all
2436 others, too. This makes the order in which console= is
2437 specified on the kernel command line less important.
2439 * libsystemd-logind gained a new sd_session_get_vt() call to
2440 retrieve the VT number of a session.
2442 * If the option "tries=0" is set for an entry of /etc/crypttab
2443 its passphrase is queried indefinitely instead of any
2444 maximum number of tries.
2446 * If a service with a configure PID file terminates its PID
2447 file will now be removed automatically if it still exists
2448 afterwards. This should put an end to stale PID files.
2450 * systemd-run will now also take relative binary path names
2451 for execution and no longer insists on absolute paths.
2453 * InaccessibleDirectories= and ReadOnlyDirectories= now take
2454 paths that are optionally prefixed with "-" to indicate that
2455 it should not be considered a failure if they do not exist.
2457 * journalctl -o (and similar commands) now understands a new
2458 output mode "short-precise", it is similar to "short" but
2459 shows timestamps with usec accuracy.
2461 * The option "discard" (as known from Debian) is now
2462 synonymous to "allow-discards" in /etc/crypttab. In fact,
2463 "discard" is preferred now (since it is easier to remember
2466 * Some licensing clean-ups were made, so that more code is now
2467 LGPL-2.1 licensed than before.
2469 * A minimal tool to save/restore the display backlight
2470 brightness across reboots has been added. It will store the
2471 backlight setting as late as possible at shutdown, and
2472 restore it as early as possible during reboot.
2474 * A logic to automatically discover and enable home and swap
2475 partitions on GPT disks has been added. With this in place
2476 /etc/fstab becomes optional for many setups as systemd can
2477 discover certain partitions located on the root disk
2478 automatically. Home partitions are recognized under their
2479 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
2480 partitions are recognized under their GPT type ID
2481 0657fd6da4ab43c484e50933c84b4f4f.
2483 * systemd will no longer pass any environment from the kernel
2484 or initrd to system services. If you want to set an
2485 environment for all services, do so via the kernel command
2486 line systemd.setenv= assignment.
2488 * The systemd-sysctl tool no longer natively reads the file
2489 /etc/sysctl.conf. If desired, the file should be symlinked
2490 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
2491 legacy support by a symlink rather than built-in code, it
2492 also makes the otherwise hidden order of application of the
2493 different files visible. (Note that this partly reverts to a
2494 pre-198 application order of sysctl knobs!)
2496 * The "systemctl set-log-level" and "systemctl dump" commands
2497 have been moved to systemd-analyze.
2499 * systemd-run learned the new --remain-after-exit switch,
2500 which causes the scope unit not to be cleaned up
2501 automatically after the process terminated.
2503 * tmpfiles learned a new --exclude-prefix= switch to exclude
2504 certain paths from operation.
2506 * journald will now automatically flush all messages to disk
2507 as soon as a message at the log level CRIT, ALERT or EMERG
2510 Contributions from: Andrew Cook, Brandon Philips, Christian
2511 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
2512 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
2513 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
2514 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
2515 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
2516 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
2517 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
2518 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
2519 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
2520 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
2521 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
2522 William Giokas, Zbigniew Jędrzejewski-Szmek
2524 -- Berlin, 2013-09-13
2528 * The documentation has been updated to cover the various new
2529 concepts introduced with 205.
2531 * Unit files now understand the new %v specifier which
2532 resolves to the kernel version string as returned by "uname
2535 * systemctl now supports filtering the unit list output by
2536 load state, active state and sub state, using the new
2539 * "systemctl status" will now show the results of the
2540 condition checks (like ConditionPathExists= and similar) of
2541 the last start attempts of the unit. They are also logged to
2544 * "journalctl -b" may now be used to look for boot output of a
2545 specific boot. Try "journalctl -b -1" for the previous boot,
2546 but the syntax is substantially more powerful.
2548 * "journalctl --show-cursor" has been added which prints the
2549 cursor string the last shown log line. This may then be used
2550 with the new "journalctl --after-cursor=" switch to continue
2551 browsing logs from that point on.
2553 * "journalctl --force" may now be used to force regeneration
2556 * Creation of "dead" device nodes has been moved from udev
2557 into kmod and tmpfiles. Previously, udev would read the kmod
2558 databases to pre-generate dead device nodes based on meta
2559 information contained in kernel modules, so that these would
2560 be auto-loaded on access rather then at boot. As this
2561 does not really have much to do with the exposing actual
2562 kernel devices to userspace this has always been slightly
2563 alien in the udev codebase. Following the new scheme kmod
2564 will now generate a runtime snippet for tmpfiles from the
2565 module meta information and it now is tmpfiles' job to the
2566 create the nodes. This also allows overriding access and
2567 other parameters for the nodes using the usual tmpfiles
2568 facilities. As side effect this allows us to remove the
2569 CAP_SYS_MKNOD capability bit from udevd entirely.
2571 * logind's device ACLs may now be applied to these "dead"
2572 devices nodes too, thus finally allowing managed access to
2573 devices such as /dev/snd/sequencer whithout loading the
2574 backing module right-away.
2576 * A new RPM macro has been added that may be used to apply
2577 tmpfiles configuration during package installation.
2579 * systemd-detect-virt and ConditionVirtualization= now can
2580 detect User-Mode-Linux machines (UML).
2582 * journald will now implicitly log the effective capabilities
2583 set of processes in the message metadata.
2585 * systemd-cryptsetup has gained support for TrueCrypt volumes.
2587 * The initrd interface has been simplified (more specifically,
2588 support for passing performance data via environment
2589 variables and fsck results via files in /run has been
2590 removed). These features were non-essential, and are
2591 nowadays available in a much nicer way by having systemd in
2592 the initrd serialize its state and have the hosts systemd
2593 deserialize it again.
2595 * The udev "keymap" data files and tools to apply keyboard
2596 specific mappings of scan to key codes, and force-release
2597 scan code lists have been entirely replaced by a udev
2598 "keyboard" builtin and a hwdb data file.
2600 * systemd will now honour the kernel's "quiet" command line
2601 argument also during late shutdown, resulting in a
2602 completely silent shutdown when used.
2604 * There's now an option to control the SO_REUSEPORT socket
2605 option in .socket units.
2607 * Instance units will now automatically get a per-template
2608 subslice of system.slice unless something else is explicitly
2609 configured. For example, instances of sshd@.service will now
2610 implicitly be placed in system-sshd.slice rather than
2611 system.slice as before.
2613 * Test coverage support may now be enabled at build time.
2615 Contributions from: Dave Reisner, Frederic Crozat, Harald
2616 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
2617 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
2618 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
2619 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
2620 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
2621 Giokas, Zbigniew Jędrzejewski-Szmek
2623 -- Berlin, 2013-07-23
2627 * Two new unit types have been introduced:
2629 Scope units are very similar to service units, however, are
2630 created out of pre-existing processes -- instead of PID 1
2631 forking off the processes. By using scope units it is
2632 possible for system services and applications to group their
2633 own child processes (worker processes) in a powerful way
2634 which then maybe used to organize them, or kill them
2635 together, or apply resource limits on them.
2637 Slice units may be used to partition system resources in an
2638 hierarchical fashion and then assign other units to them. By
2639 default there are now three slices: system.slice (for all
2640 system services), user.slice (for all user sessions),
2641 machine.slice (for VMs and containers).
2643 Slices and scopes have been introduced primarily in
2644 context of the work to move cgroup handling to a
2645 single-writer scheme, where only PID 1
2646 creates/removes/manages cgroups.
2648 * There's a new concept of "transient" units. In contrast to
2649 normal units these units are created via an API at runtime,
2650 not from configuration from disk. More specifically this
2651 means it is now possible to run arbitrary programs as
2652 independent services, with all execution parameters passed
2653 in via bus APIs rather than read from disk. Transient units
2654 make systemd substantially more dynamic then it ever was,
2655 and useful as a general batch manager.
2657 * logind has been updated to make use of scope and slice units
2658 for managing user sessions. As a user logs in he will get
2659 his own private slice unit, to which all sessions are added
2660 as scope units. We also added support for automatically
2661 adding an instance of user@.service for the user into the
2662 slice. Effectively logind will no longer create cgroup
2663 hierarchies on its own now, it will defer entirely to PID 1
2664 for this by means of scope, service and slice units. Since
2665 user sessions this way become entities managed by PID 1
2666 the output of "systemctl" is now a lot more comprehensive.
2668 * A new mini-daemon "systemd-machined" has been added which
2669 may be used by virtualization managers to register local
2670 VMs/containers. nspawn has been updated accordingly, and
2671 libvirt will be updated shortly. machined will collect a bit
2672 of meta information about the VMs/containers, and assign
2673 them their own scope unit (see above). The collected
2674 meta-data is then made available via the "machinectl" tool,
2675 and exposed in "ps" and similar tools. machined/machinectl
2676 is compile-time optional.
2678 * As discussed earlier, the low-level cgroup configuration
2679 options ControlGroup=, ControlGroupModify=,
2680 ControlGroupPersistent=, ControlGroupAttribute= have been
2681 removed. Please use high-level attribute settings instead as
2682 well as slice units.
2684 * A new bus call SetUnitProperties() has been added to alter
2685 various runtime parameters of a unit. This is primarily
2686 useful to alter cgroup parameters dynamically in a nice way,
2687 but will be extended later on to make more properties
2688 modifiable at runtime. systemctl gained a new set-properties
2689 command that wraps this call.
2691 * A new tool "systemd-run" has been added which can be used to
2692 run arbitrary command lines as transient services or scopes,
2693 while configuring a number of settings via the command
2694 line. This tool is currently very basic, however already
2695 very useful. We plan to extend this tool to even allow
2696 queuing of execution jobs with time triggers from the
2697 command line, similar in fashion to "at".
2699 * nspawn will now inform the user explicitly that kernels with
2700 audit enabled break containers, and suggest the user to turn
2703 * Support for detecting the IMA and AppArmor security
2704 frameworks with ConditionSecurity= has been added.
2706 * journalctl gained a new "-k" switch for showing only kernel
2707 messages, mimicking dmesg output; in addition to "--user"
2708 and "--system" switches for showing only user's own logs
2711 * systemd-delta can now show information about drop-in
2712 snippets extending unit files.
2714 * libsystemd-bus has been substantially updated but is still
2715 not available as public API.
2717 * systemd will now look for the "debug" argument on the kernel
2718 command line and enable debug logging, similar to what
2719 "systemd.log_level=debug" already did before.
2721 * "systemctl set-default", "systemctl get-default" has been
2722 added to configure the default.target symlink, which
2723 controls what to boot into by default.
2725 * "systemctl set-log-level" has been added as a convenient
2726 way to raise and lower systemd logging threshold.
2728 * "systemd-analyze plot" will now show the time the various
2729 generators needed for execution, as well as information
2730 about the unit file loading.
2732 * libsystemd-journal gained a new sd_journal_open_files() call
2733 for opening specific journal files. journactl also gained a
2734 new switch to expose this new functionality. Previously we
2735 only supported opening all files from a directory, or all
2736 files from the system, as opening individual files only is
2737 racy due to journal file rotation.
2739 * systemd gained the new DefaultEnvironment= setting in
2740 /etc/systemd/system.conf to set environment variables for
2743 * If a privileged process logs a journal message with the
2744 OBJECT_PID= field set, then journald will automatically
2745 augment this with additional OBJECT_UID=, OBJECT_GID=,
2746 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
2747 system services want to log events about specific client
2748 processes. journactl/systemctl has been updated to make use
2749 of this information if all log messages regarding a specific
2752 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
2753 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
2754 Reisner, David Coppa, David King, David Strauss, Eelco
2755 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
2756 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
2757 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
2758 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
2759 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
2760 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
2761 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
2762 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
2763 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
2764 Łukasz Stelmach, 장동준
2768 * The Python bindings gained some minimal support for the APIs
2769 exposed by libsystemd-logind.
2771 * ConditionSecurity= gained support for detecting SMACK. Since
2772 this condition already supports SELinux and AppArmor we only
2773 miss IMA for this. Patches welcome!
2775 Contributions from: Karol Lewandowski, Lennart Poettering,
2776 Zbigniew Jędrzejewski-Szmek
2780 * systemd-nspawn will now create /etc/resolv.conf if
2781 necessary, before bind-mounting the host's file onto it.
2783 * systemd-nspawn will now store meta information about a
2784 container on the container's cgroup as extended attribute
2785 fields, including the root directory.
2787 * The cgroup hierarchy has been reworked in many ways. All
2788 objects any of the components systemd creates in the cgroup
2789 tree are now suffixed. More specifically, user sessions are
2790 now placed in cgroups suffixed with ".session", users in
2791 cgroups suffixed with ".user", and nspawn containers in
2792 cgroups suffixed with ".nspawn". Furthermore, all cgroup
2793 names are now escaped in a simple scheme to avoid collision
2794 of userspace object names with kernel filenames. This work
2795 is preparation for making these objects relocatable in the
2796 cgroup tree, in order to allow easy resource partitioning of
2797 these objects without causing naming conflicts.
2799 * systemctl list-dependencies gained the new switches
2800 --plain, --reverse, --after and --before.
2802 * systemd-inhibit now shows the process name of processes that
2803 have taken an inhibitor lock.
2805 * nss-myhostname will now also resolve "localhost"
2806 implicitly. This makes /etc/hosts an optional file and
2807 nicely handles that on IPv6 ::1 maps to both "localhost" and
2810 * libsystemd-logind.so gained a new call
2811 sd_get_machine_names() to enumerate running containers and
2812 VMs (currently only supported by very new libvirt and
2813 nspawn). sd_login_monitor can now be used to watch
2814 VMs/containers coming and going.
2816 * .include is not allowed recursively anymore, and only in
2817 unit files. Usually it is better to use drop-in snippets in
2818 .d/*.conf anyway, as introduced with systemd 198.
2820 * systemd-analyze gained a new "critical-chain" command that
2821 determines the slowest chain of units run during system
2822 boot-up. It is very useful for tracking down where
2823 optimizing boot time is the most beneficial.
2825 * systemd will no longer allow manipulating service paths in
2826 the name=systemd:/system cgroup tree using ControlGroup= in
2827 units. (But is still fine with it in all other dirs.)
2829 * There's a new systemd-nspawn@.service service file that may
2830 be used to easily run nspawn containers as system
2831 services. With the container's root directory in
2832 /var/lib/container/foobar it is now sufficient to run
2833 "systemctl start systemd-nspawn@foobar.service" to boot it.
2835 * systemd-cgls gained a new parameter "--machine" to list only
2836 the processes within a certain container.
2838 * ConditionSecurity= now can check for "apparmor". We still
2839 are lacking checks for SMACK and IMA for this condition
2840 check though. Patches welcome!
2842 * A new configuration file /etc/systemd/sleep.conf has been
2843 added that may be used to configure which kernel operation
2844 systemd is supposed to execute when "suspend", "hibernate"
2845 or "hybrid-sleep" is requested. This makes the new kernel
2846 "freeze" state accessible to the user.
2848 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
2849 the passed argument if applicable.
2851 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
2852 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
2853 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
2854 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
2855 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
2856 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
2857 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
2862 * The output of 'systemctl list-jobs' got some polishing. The
2863 '--type=' argument may now be passed more than once. A new
2864 command 'systemctl list-sockets' has been added which shows
2865 a list of kernel sockets systemd is listening on with the
2866 socket units they belong to, plus the units these socket
2869 * The experimental libsystemd-bus library got substantial
2870 updates to work in conjunction with the (also experimental)
2871 kdbus kernel project. It works well enough to exchange
2872 messages with some sophistication. Note that kdbus is not
2873 ready yet, and the library is mostly an elaborate test case
2874 for now, and not installable.
2876 * systemd gained a new unit 'systemd-static-nodes.service'
2877 that generates static device nodes earlier during boot, and
2878 can run in conjunction with udev.
2880 * libsystemd-login gained a new call sd_pid_get_user_unit()
2881 to retrieve the user systemd unit a process is running
2882 in. This is useful for systems where systemd is used as
2885 * systemd-nspawn now places all containers in the new /machine
2886 top-level cgroup directory in the name=systemd
2887 hierarchy. libvirt will soon do the same, so that we get a
2888 uniform separation of /system, /user and /machine for system
2889 services, user processes and containers/virtual
2890 machines. This new cgroup hierarchy is also useful to stick
2891 stable names to specific container instances, which can be
2892 recognized later this way (this name may be controlled
2893 via systemd-nspawn's new -M switch). libsystemd-login also
2894 gained a new call sd_pid_get_machine_name() to retrieve the
2895 name of the container/VM a specific process belongs to.
2897 * bootchart can now store its data in the journal.
2899 * libsystemd-journal gained a new call
2900 sd_journal_add_conjunction() for AND expressions to the
2901 matching logic. This can be used to express more complex
2902 logical expressions.
2904 * journactl can now take multiple --unit= and --user-unit=
2907 * The cryptsetup logic now understands the "luks.key=" kernel
2908 command line switch for specifying a file to read the
2909 decryption key from. Also, if a configured key file is not
2910 found the tool will now automatically fall back to prompting
2913 * Python systemd.journal module was updated to wrap recently
2914 added functions from libsystemd-journal. The interface was
2915 changed to bring the low level interface in s.j._Reader
2916 closer to the C API, and the high level interface in
2917 s.j.Reader was updated to wrap and convert all data about
2920 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
2921 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
2922 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
2923 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
2924 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
2925 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2929 * journalctl --update-catalog now understands a new --root=
2930 option to operate on catalogs found in a different root
2933 * During shutdown after systemd has terminated all running
2934 services a final killing loop kills all remaining left-over
2935 processes. We will now print the name of these processes
2936 when we send SIGKILL to them, since this usually indicates a
2939 * If /etc/crypttab refers to password files stored on
2940 configured mount points automatic dependencies will now be
2941 generated to ensure the specific mount is established first
2942 before the key file is attempted to be read.
2944 * 'systemctl status' will now show information about the
2945 network sockets a socket unit is listening on.
2947 * 'systemctl status' will also shown information about any
2948 drop-in configuration file for units. (Drop-In configuration
2949 files in this context are files such as
2950 /etc/systemd/systemd/foobar.service.d/*.conf)
2952 * systemd-cgtop now optionally shows summed up CPU times of
2953 cgroups. Press '%' while running cgtop to switch between
2954 percentage and absolute mode. This is useful to determine
2955 which cgroups use up the most CPU time over the entire
2956 runtime of the system. systemd-cgtop has also been updated
2957 to be 'pipeable' for processing with further shell tools.
2959 * 'hostnamectl set-hostname' will now allow setting of FQDN
2962 * The formatting and parsing of time span values has been
2963 changed. The parser now understands fractional expressions
2964 such as "5.5h". The formatter will now output fractional
2965 expressions for all time spans under 1min, i.e. "5.123456s"
2966 rather than "5s 123ms 456us". For time spans under 1s
2967 millisecond values are shown, for those under 1ms
2968 microsecond values are shown. This should greatly improve
2969 all time-related output of systemd.
2971 * libsystemd-login and libsystemd-journal gained new
2972 functions for querying the poll() events mask and poll()
2973 timeout value for integration into arbitrary event
2976 * localectl gained the ability to list available X11 keymaps
2977 (models, layouts, variants, options).
2979 * 'systemd-analyze dot' gained the ability to filter for
2980 specific units via shell-style globs, to create smaller,
2981 more useful graphs. I.e. it is now possible to create simple
2982 graphs of all the dependencies between only target units, or
2983 of all units that Avahi has dependencies with.
2985 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
2986 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
2987 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
2988 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
2989 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
2990 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
2991 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
2995 * The boot-time readahead implementation for rotating media
2996 will now read the read-ahead data in multiple passes which
2997 consist of all read requests made in equidistant time
2998 intervals. This means instead of strictly reading read-ahead
2999 data in its physical order on disk we now try to find a
3000 middle ground between physical and access time order.
3002 * /etc/os-release files gained a new BUILD_ID= field for usage
3003 on operating systems that provide continuous builds of OS
3006 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
3007 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
3008 William Douglas, Zbigniew Jędrzejewski-Szmek
3012 * systemd-python gained an API exposing libsystemd-daemon.
3014 * The SMACK setup logic gained support for uploading CIPSO
3017 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
3018 ReadOnlyDirectories= and InaccessibleDirectories= has
3019 changed. The private /tmp and /var/tmp directories are now
3020 shared by all processes of a service (which means
3021 ExecStartPre= may now leave data in /tmp that ExecStart= of
3022 the same service can still access). When a service is
3023 stopped its temporary directories are immediately deleted
3024 (normal clean-up with tmpfiles is still done in addition to
3027 * By default, systemd will now set a couple of sysctl
3028 variables in the kernel: the safe sysrq options are turned
3029 on, IP route verification is turned on, and source routing
3030 disabled. The recently added hardlink and softlink
3031 protection of the kernel is turned on. These settings should
3032 be reasonably safe, and good defaults for all new systems.
3034 * The predictable network naming logic may now be turned off
3035 with a new kernel command line switch: net.ifnames=0.
3037 * A new libsystemd-bus module has been added that implements a
3038 pretty complete D-Bus client library. For details see:
3040 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
3042 * journald will now explicitly flush the journal files to disk
3043 at the latest 5min after each write. The file will then also
3044 be marked offline until the next write. This should increase
3045 reliability in case of a crash. The synchronization delay
3046 can be configured via SyncIntervalSec= in journald.conf.
3048 * There's a new remote-fs-setup.target unit that can be used
3049 to pull in specific services when at least one remote file
3050 system is to be mounted.
3052 * There are new targets timers.target and paths.target as
3053 canonical targets to pull user timer and path units in
3054 from. This complements sockets.target with a similar
3055 purpose for socket units.
3057 * libudev gained a new call udev_device_set_attribute_value()
3058 to set sysfs attributes of a device.
3060 * The udev daemon now sets the default number of worker
3061 processes executed in parallel based on the number of available
3062 CPUs instead of the amount of available RAM. This is supposed
3063 to provide a more reliable default and limit a too aggressive
3064 paralellism for setups with 1000s of devices connected.
3066 Contributions from: Auke Kok, Colin Walters, Cristian
3067 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
3068 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
3069 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
3070 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
3071 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
3072 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
3073 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
3074 Zbigniew Jędrzejewski-Szmek
3078 * Configuration of unit files may now be extended via drop-in
3079 files without having to edit/override the unit files
3080 themselves. More specifically, if the administrator wants to
3081 change one value for a service file foobar.service he can
3082 now do so by dropping in a configuration snippet into
3083 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
3084 will load all these snippets and apply them on top of the
3085 main unit configuration file, possibly extending or
3086 overriding its settings. Using these drop-in snippets is
3087 generally nicer than the two earlier options for changing
3088 unit files locally: copying the files from
3089 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
3090 them there; or creating a new file in /etc/systemd/system/
3091 that incorporates the original one via ".include". Drop-in
3092 snippets into these .d/ directories can be placed in any
3093 directory systemd looks for units in, and the usual
3094 overriding semantics between /usr/lib, /etc and /run apply
3097 * Most unit file settings which take lists of items can now be
3098 reset by assigning the empty string to them. For example,
3099 normally, settings such as Environment=FOO=BAR append a new
3100 environment variable assignment to the environment block,
3101 each time they are used. By assigning Environment= the empty
3102 string the environment block can be reset to empty. This is
3103 particularly useful with the .d/*.conf drop-in snippets
3104 mentioned above, since this adds the ability to reset list
3105 settings from vendor unit files via these drop-ins.
3107 * systemctl gained a new "list-dependencies" command for
3108 listing the dependencies of a unit recursively.
3110 * Inhibitors are now honored and listed by "systemctl
3111 suspend", "systemctl poweroff" (and similar) too, not only
3112 GNOME. These commands will also list active sessions by
3115 * Resource limits (as exposed by the various control group
3116 controllers) can now be controlled dynamically at runtime
3117 for all units. More specifically, you can now use a command
3118 like "systemctl set-cgroup-attr foobar.service cpu.shares
3119 2000" to alter the CPU shares a specific service gets. These
3120 settings are stored persistently on disk, and thus allow the
3121 administrator to easily adjust the resource usage of
3122 services with a few simple commands. This dynamic resource
3123 management logic is also available to other programs via the
3124 bus. Almost any kernel cgroup attribute and controller is
3127 * systemd-vconsole-setup will now copy all font settings to
3128 all allocated VTs, where it previously applied them only to
3131 * libsystemd-login gained the new sd_session_get_tty() API
3134 * This release drops support for a few legacy or
3135 distribution-specific LSB facility names when parsing init
3136 scripts: $x-display-manager, $mail-transfer-agent,
3137 $mail-transport-agent, $mail-transfer-agent, $smtp,
3138 $null. Also, the mail-transfer-agent.target unit backing
3139 this has been removed. Distributions which want to retain
3140 compatibility with this should carry the burden for
3141 supporting this themselves and patch support for these back
3142 in, if they really need to. Also, the facilities $syslog and
3143 $local_fs are now ignored, since systemd does not support
3144 early-boot LSB init scripts anymore, and these facilities
3145 are implied anyway for normal services. syslog.target has
3148 * There are new bus calls on PID1's Manager object for
3149 cancelling jobs, and removing snapshot units. Previously,
3150 both calls were only available on the Job and Snapshot
3153 * systemd-journal-gatewayd gained SSL support.
3155 * The various "environment" files, such as /etc/locale.conf
3156 now support continuation lines with a backslash ("\") as
3157 last character in the line, similarly in style (but different)
3158 to how this is supported in shells.
3160 * For normal user processes the _SYSTEMD_USER_UNIT= field is
3161 now implicitly appended to every log entry logged. systemctl
3162 has been updated to filter by this field when operating on a
3163 user systemd instance.
3165 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
3166 CAP_AUDIT_CONTROL capabilities to the capabilities set for
3167 the container. This makes it easier to boot unmodified
3168 Fedora systems in a container, which however still requires
3169 audit=0 to be passed on the kernel command line. Auditing in
3170 kernel and userspace is unfortunately still too broken in
3171 context of containers, hence we recommend compiling it out
3172 of the kernel or using audit=0. Hopefully this will be fixed
3173 one day for good in the kernel.
3175 * nspawn gained the new --bind= and --bind-ro= parameters to
3176 bind mount specific directories from the host into the
3179 * nspawn will now mount its own devpts file system instance
3180 into the container, in order not to leak pty devices from
3181 the host into the container.
3183 * systemd will now read the firmware boot time performance
3184 information from the EFI variables, if the used boot loader
3185 supports this, and takes it into account for boot performance
3186 analysis via "systemd-analyze". This is currently supported
3187 only in conjunction with Gummiboot, but could be supported
3188 by other boot loaders too. For details see:
3190 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
3192 * A new generator has been added that automatically mounts the
3193 EFI System Partition (ESP) to /boot, if that directory
3194 exists, is empty, and no other file system has been
3195 configured to be mounted there.
3197 * logind will now send out PrepareForSleep(false) out
3198 unconditionally, after coming back from suspend. This may be
3199 used by applications as asynchronous notification for
3200 system resume events.
3202 * "systemctl unlock-sessions" has been added, that allows
3203 unlocking the screens of all user sessions at once, similar
3204 to how "systemctl lock-sessions" already locked all users
3205 sessions. This is backed by a new D-Bus call UnlockSessions().
3207 * "loginctl seat-status" will now show the master device of a
3208 seat. (i.e. the device of a seat that needs to be around for
3209 the seat to be considered available, usually the graphics
3212 * tmpfiles gained a new "X" line type, that allows
3213 configuration of files and directories (with wildcards) that
3214 shall be excluded from automatic cleanup ("aging").
3216 * udev default rules set the device node permissions now only
3217 at "add" events, and do not change them any longer with a
3218 later "change" event.
3220 * The log messages for lid events and power/sleep keypresses
3221 now carry a message ID.
3223 * We now have a substantially larger unit test suite, but this
3224 continues to be work in progress.
3226 * udevadm hwdb gained a new --root= parameter to change the
3227 root directory to operate relative to.
3229 * logind will now issue a background sync() request to the kernel
3230 early at shutdown, so that dirty buffers are flushed to disk early
3231 instead of at the last moment, in order to optimize shutdown
3234 * A new bootctl tool has been added that is an interface for
3235 certain boot loader operations. This is currently a preview
3236 and is likely to be extended into a small mechanism daemon
3237 like timedated, localed, hostnamed, and can be used by
3238 graphical UIs to enumerate available boot options, and
3239 request boot into firmware operations.
3241 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
3242 the rest of the package. It also has been updated to work
3243 correctly in initrds.
3245 * Policykit previously has been runtime optional, and is now
3246 also compile time optional via a configure switch.
3248 * systemd-analyze has been reimplemented in C. Also "systemctl
3249 dot" has moved into systemd-analyze.
3251 * "systemctl status" with no further parameters will now print
3252 the status of all active or failed units.
3254 * Operations such as "systemctl start" can now be executed
3255 with a new mode "--irreversible" which may be used to queue
3256 operations that cannot accidentally be reversed by a later
3257 job queuing. This is by default used to make shutdown
3258 requests more robust.
3260 * The Python API of systemd now gained a new module for
3261 reading journal files.
3263 * A new tool kernel-install has been added that can install
3264 kernel images according to the Boot Loader Specification:
3266 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
3268 * Boot time console output has been improved to provide
3269 animated boot time output for hanging jobs.
3271 * A new tool systemd-activate has been added which can be used
3272 to test socket activation with, directly from the command
3273 line. This should make it much easier to test and debug
3274 socket activation in daemons.
3276 * journalctl gained a new "--reverse" (or -r) option to show
3277 journal output in reverse order (i.e. newest line first).
3279 * journalctl gained a new "--pager-end" (or -e) option to jump
3280 to immediately jump to the end of the journal in the
3281 pager. This is only supported in conjunction with "less".
3283 * journalctl gained a new "--user-unit=" option, that works
3284 similarly to "--unit=" but filters for user units rather than
3287 * A number of unit files to ease adoption of systemd in
3288 initrds has been added. This moves some minimal logic from
3289 the various initrd implementations into systemd proper.
3291 * The journal files are now owned by a new group
3292 "systemd-journal", which exists specifically to allow access
3293 to the journal, and nothing else. Previously, we used the
3294 "adm" group for that, which however possibly covers more
3295 than just journal/log file access. This new group is now
3296 already used by systemd-journal-gatewayd to ensure this
3297 daemon gets access to the journal files and as little else
3298 as possible. Note that "make install" will also set FS ACLs
3299 up for /var/log/journal to give "adm" and "wheel" read
3300 access to it, in addition to "systemd-journal" which owns
3301 the journal files. We recommend that packaging scripts also
3302 add read access to "adm" + "wheel" to /var/log/journal, and
3303 all existing/future journal files. To normal users and
3304 administrators little changes, however packagers need to
3305 ensure to create the "systemd-journal" system group at
3306 package installation time.
3308 * The systemd-journal-gatewayd now runs as unprivileged user
3309 systemd-journal-gateway:systemd-journal-gateway. Packaging
3310 scripts need to create these system user/group at
3313 * timedated now exposes a new boolean property CanNTP that
3314 indicates whether a local NTP service is available or not.
3316 * systemd-detect-virt will now also detect xen PVs
3318 * The pstore file system is now mounted by default, if it is
3321 * In addition to the SELinux and IMA policies we will now also
3322 load SMACK policies at early boot.
3324 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
3325 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
3326 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
3327 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
3328 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
3329 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
3330 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
3331 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
3332 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
3333 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
3334 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
3335 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
3336 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
3337 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
3341 * Timer units now support calendar time events in addition to
3342 monotonic time events. That means you can now trigger a unit
3343 based on a calendar time specification such as "Thu,Fri
3344 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
3345 or fifth day of any month of the year 2013, given that it is
3346 a thursday or friday. This brings timer event support
3347 considerably closer to cron's capabilities. For details on
3348 the supported calendar time specification language see
3351 * udev now supports a number of different naming policies for
3352 network interfaces for predictable names, and a combination
3353 of these policies is now the default. Please see this wiki
3354 document for details:
3356 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
3358 * Auke Kok's bootchart implementation has been added to the
3359 systemd tree. It is an optional component that can graph the
3360 boot in quite some detail. It is one of the best bootchart
3361 implementations around and minimal in its code and
3364 * nss-myhostname has been integrated into the systemd source
3365 tree. nss-myhostname guarantees that the local hostname
3366 always stays resolvable via NSS. It has been a weak
3367 requirement of systemd-hostnamed since a long time, and
3368 since its code is actually trivial we decided to just
3369 include it in systemd's source tree. It can be turned off
3370 with a configure switch.
3372 * The read-ahead logic is now capable of properly detecting
3373 whether a btrfs file system is on SSD or rotating media, in
3374 order to optimize the read-ahead scheme. Previously, it was
3375 only capable of detecting this on traditional file systems
3378 * In udev, additional device properties are now read from the
3379 IAB in addition to the OUI database. Also, Bluetooth company
3380 identities are attached to the devices as well.
3382 * In service files %U may be used as specifier that is
3383 replaced by the configured user name of the service.
3385 * nspawn may now be invoked without a controlling TTY. This
3386 makes it suitable for invocation as its own service. This
3387 may be used to set up a simple containerized server system
3388 using only core OS tools.
3390 * systemd and nspawn can now accept socket file descriptors
3391 when they are started for socket activation. This enables
3392 implementation of socket activated nspawn
3393 containers. i.e. think about autospawning an entire OS image
3394 when the first SSH or HTTP connection is received. We expect
3395 that similar functionality will also be added to libvirt-lxc
3398 * journalctl will now suppress ANSI color codes when
3399 presenting log data.
3401 * systemctl will no longer show control group information for
3402 a unit if a the control group is empty anyway.
3404 * logind can now automatically suspend/hibernate/shutdown the
3407 * /etc/machine-info and hostnamed now also expose the chassis
3408 type of the system. This can be used to determine whether
3409 the local system is a laptop, desktop, handset or
3410 tablet. This information may either be configured by the
3411 user/vendor or is automatically determined from ACPI and DMI
3412 information if possible.
3414 * A number of PolicyKit actions are now bound together with
3415 "imply" rules. This should simplify creating UIs because
3416 many actions will now authenticate similar ones as well.
3418 * Unit files learnt a new condition ConditionACPower= which
3419 may be used to conditionalize a unit depending on whether an
3420 AC power source is connected or not, of whether the system
3421 is running on battery power.
3423 * systemctl gained a new "is-failed" verb that may be used in
3424 shell scripts and suchlike to check whether a specific unit
3425 is in the "failed" state.
3427 * The EnvironmentFile= setting in unit files now supports file
3428 globbing, and can hence be used to easily read a number of
3429 environment files at once.
3431 * systemd will no longer detect and recognize specific
3432 distributions. All distribution-specific #ifdeffery has been
3433 removed, systemd is now fully generic and
3434 distribution-agnostic. Effectively, not too much is lost as
3435 a lot of the code is still accessible via explicit configure
3436 switches. However, support for some distribution specific
3437 legacy configuration file formats has been dropped. We
3438 recommend distributions to simply adopt the configuration
3439 files everybody else uses now and convert the old
3440 configuration from packaging scripts. Most distributions
3441 already did that. If that's not possible or desirable,
3442 distributions are welcome to forward port the specific
3443 pieces of code locally from the git history.
3445 * When logging a message about a unit systemd will now always
3446 log the unit name in the message meta data.
3448 * localectl will now also discover system locale data that is
3449 not stored in locale archives, but directly unpacked.
3451 * logind will no longer unconditionally use framebuffer
3452 devices as seat masters, i.e. as devices that are required
3453 to be existing before a seat is considered preset. Instead,
3454 it will now look for all devices that are tagged as
3455 "seat-master" in udev. By default framebuffer devices will
3456 be marked as such, but depending on local systems other
3457 devices might be marked as well. This may be used to
3458 integrate graphics cards using closed source drivers (such
3459 as NVidia ones) more nicely into logind. Note however, that
3460 we recommend using the open source NVidia drivers instead,
3461 and no udev rules for the closed-source drivers will be
3462 shipped from us upstream.
3464 Contributions from: Adam Williamson, Alessandro Crismani, Auke
3465 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
3466 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
3467 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
3468 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
3469 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
3470 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
3471 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
3472 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
3473 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
3474 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
3475 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
3480 * udev gained support for loading additional device properties
3481 from an indexed database that is keyed by vendor/product IDs
3482 and similar device identifiers. For the beginning this
3483 "hwdb" is populated with data from the well-known PCI and
3484 USB database, but also includes PNP, ACPI and OID data. In
3485 the longer run this indexed database shall grow into
3486 becoming the one central database for non-essential
3487 userspace device metadata. Previously, data from the PCI/USB
3488 database was only attached to select devices, since the
3489 lookup was a relatively expensive operation due to O(n) time
3490 complexity (with n being the number of entries in the
3491 database). Since this is now O(1), we decided to add in this
3492 data for all devices where this is available, by
3493 default. Note that the indexed database needs to be rebuilt
3494 when new data files are installed. To achieve this you need
3495 to update your packaging scripts to invoke "udevadm hwdb
3496 --update" after installation of hwdb data files. For
3497 RPM-based distributions we introduced the new
3498 %udev_hwdb_update macro for this purpose.
3500 * The Journal gained support for the "Message Catalog", an
3501 indexed database to link up additional information with
3502 journal entries. For further details please check:
3504 http://www.freedesktop.org/wiki/Software/systemd/catalog
3506 The indexed message catalog database also needs to be
3507 rebuilt after installation of message catalog files. Use
3508 "journalctl --update-catalog" for this. For RPM-based
3509 distributions we introduced the %journal_catalog_update
3510 macro for this purpose.
3512 * The Python Journal bindings gained support for the standard
3513 Python logging framework.
3515 * The Journal API gained new functions for checking whether
3516 the underlying file system of a journal file is capable of
3517 properly reporting file change notifications, or whether
3518 applications that want to reflect journal changes "live"
3519 need to recheck journal files continuously in appropriate
3522 * It is now possible to set the "age" field for tmpfiles
3523 entries to 0, indicating that files matching this entry
3524 shall always be removed when the directories are cleaned up.
3526 * coredumpctl gained a new "gdb" verb which invokes gdb
3527 right-away on the selected coredump.
3529 * There's now support for "hybrid sleep" on kernels that
3530 support this, in addition to "suspend" and "hibernate". Use
3531 "systemctl hybrid-sleep" to make use of this.
3533 * logind's HandleSuspendKey= setting (and related settings)
3534 now gained support for a new "lock" setting to simply
3535 request the screen lock on all local sessions, instead of
3536 actually executing a suspend or hibernation.
3538 * systemd will now mount the EFI variables file system by
3541 * Socket units now gained support for configuration of the
3542 SMACK security label.
3544 * timedatectl will now output the time of the last and next
3545 daylight saving change.
3547 * We dropped support for various legacy and distro-specific
3548 concepts, such as insserv, early-boot SysV services
3549 (i.e. those for non-standard runlevels such as 'b' or 'S')
3550 or ArchLinux /etc/rc.conf support. We recommend the
3551 distributions who still need support this to either continue
3552 to maintain the necessary patches downstream, or find a
3553 different solution. (Talk to us if you have questions!)
3555 * Various systemd components will now bypass PolicyKit checks
3556 for root and otherwise handle properly if PolicyKit is not
3557 found to be around. This should fix most issues for
3558 PolicyKit-less systems. Quite frankly this should have been
3559 this way since day one. It is absolutely our intention to
3560 make systemd work fine on PolicyKit-less systems, and we
3561 consider it a bug if something does not work as it should if
3562 PolicyKit is not around.
3564 * For embedded systems it is now possible to build udev and
3565 systemd without blkid and/or kmod support.
3567 * "systemctl switch-root" is now capable of switching root
3568 more than once. I.e. in addition to transitions from the
3569 initrd to the host OS it is now possible to transition to
3570 further OS images from the host. This is useful to implement
3571 offline updating tools.
3573 * Various other additions have been made to the RPM macros
3574 shipped with systemd. Use %udev_rules_update() after
3575 installing new udev rules files. %_udevhwdbdir,
3576 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
3577 %_sysctldir are now available which resolve to the right
3578 directories for packages to place various data files in.
3580 * journalctl gained the new --full switch (in addition to
3581 --all, to disable ellipsation for long messages.
3583 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
3584 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
3585 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
3586 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
3587 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
3588 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
3589 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
3590 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
3591 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
3595 * journalctl gained new --since= and --until= switches to
3596 filter by time. It also now supports nice filtering for
3597 units via --unit=/-u.
3599 * Type=oneshot services may use ExecReload= and do the
3602 * The journal daemon now supports time-based rotation and
3603 vacuuming, in addition to the usual disk-space based
3606 * The journal will now index the available field values for
3607 each field name. This enables clients to show pretty drop
3608 downs of available match values when filtering. The bash
3609 completion of journalctl has been updated
3610 accordingly. journalctl gained a new switch -F to list all
3611 values a certain field takes in the journal database.
3613 * More service events are now written as structured messages
3614 to the journal, and made recognizable via message IDs.
3616 * The timedated, localed and hostnamed mini-services which
3617 previously only provided support for changing time, locale
3618 and hostname settings from graphical DEs such as GNOME now
3619 also have a minimal (but very useful) text-based client
3620 utility each. This is probably the nicest way to changing
3621 these settings from the command line now, especially since
3622 it lists available options and is fully integrated with bash
3625 * There's now a new tool "systemd-coredumpctl" to list and
3626 extract coredumps from the journal.
3628 * We now install a README each in /var/log/ and
3629 /etc/rc.d/init.d explaining where the system logs and init
3630 scripts went. This hopefully should help folks who go to
3631 that dirs and look into the otherwise now empty void and
3632 scratch their heads.
3634 * When user-services are invoked (by systemd --user) the
3635 $MANAGERPID env var is set to the PID of systemd.
3637 * SIGRTMIN+24 when sent to a --user instance will now result
3638 in immediate termination of systemd.
3640 * gatewayd received numerous feature additions such as a
3641 "follow" mode, for live syncing and filtering.
3643 * browse.html now allows filtering and showing detailed
3644 information on specific entries. Keyboard navigation and
3645 mouse screen support has been added.
3647 * gatewayd/journalctl now supports HTML5/JSON
3648 Server-Sent-Events as output.
3650 * The SysV init script compatibility logic will now
3651 heuristically determine whether a script supports the
3652 "reload" verb, and only then make this available as
3655 * "systemctl status --follow" has been removed, use "journalctl
3658 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
3659 have been removed since they are hardly useful to be
3662 * And I'd like to take the opportunity to specifically mention
3663 Zbigniew for his great contributions. Zbigniew, you rock!
3665 Contributions from: Andrew Eikum, Christian Hesse, Colin
3666 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
3667 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
3668 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
3669 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
3670 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
3671 Jędrzejewski-Szmek, Сковорода Никита Андреевич
3675 * If /etc/vconsole.conf is non-existent or empty we will no
3676 longer load any console font or key map at boot by
3677 default. Instead the kernel defaults will be left
3678 intact. This is definitely the right thing to do, as no
3679 configuration should mean no configuration, and hard-coding
3680 font names that are different on all archs is probably a bad
3681 idea. Also, the kernel default key map and font should be
3682 good enough for most cases anyway, and mostly identical to
3683 the userspace fonts/key maps we previously overloaded them
3684 with. If distributions want to continue to default to a
3685 non-kernel font or key map they should ship a default
3686 /etc/vconsole.conf with the appropriate contents.
3688 Contributions from: Colin Walters, Daniel J Walsh, Dave
3689 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
3690 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
3694 * journalctl gained a new --cursor= switch to show entries
3695 starting from the specified location in the journal.
3697 * We now enforce a size limit on journal entry fields exported
3698 with "-o json" in journalctl. Fields larger than 4K will be
3699 assigned null. This can be turned off with --all.
3701 * An (optional) journal gateway daemon is now available as
3702 "systemd-journal-gatewayd.service". This service provides
3703 access to the journal via HTTP and JSON. This functionality
3704 will be used to implement live log synchronization in both
3705 pull and push modes, but has various other users too, such
3706 as easy log access for debugging of embedded devices. Right
3707 now it is already useful to retrieve the journal via HTTP:
3709 # systemctl start systemd-journal-gatewayd.service
3710 # wget http://localhost:19531/entries
3712 This will download the journal contents in a
3713 /var/log/messages compatible format. The same as JSON:
3715 # curl -H"Accept: application/json" http://localhost:19531/entries
3717 This service is also accessible via a web browser where a
3718 single static HTML5 app is served that uses the JSON logic
3719 to enable the user to do some basic browsing of the
3720 journal. This will be extended later on. Here's an example
3721 screenshot of this app in its current state:
3723 http://0pointer.de/public/journal-gatewayd
3725 Contributions from: Kay Sievers, Lennart Poettering, Robert
3726 Milasan, Tom Gundersen
3730 * The bash completion logic is now available for journalctl
3733 * We do not mount the "cpuset" controller anymore together with
3734 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
3735 started if no parameters are assigned to it. "cpuset" hence
3736 broke code that assumed it it could create "cpu" groups and
3739 * journalctl -f will now subscribe to terminal size changes,
3740 and line break accordingly.
3742 Contributions from: Dave Reisner, Kay Sievers, Lennart
3743 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
3747 * nspawn will now create a symlink /etc/localtime in the
3748 container environment, copying the host's timezone
3749 setting. Previously this has been done via a bind mount, but
3750 since symlinks cannot be bind mounted this has now been
3751 changed to create/update the appropriate symlink.
3753 * journalctl -n's line number argument is now optional, and
3754 will default to 10 if omitted.
3756 * journald will now log the maximum size the journal files may
3757 take up on disk. This is particularly useful if the default
3758 built-in logic of determining this parameter from the file
3759 system size is used. Use "systemctl status
3760 systemd-journald.service" to see this information.
3762 * The multi-seat X wrapper tool has been stripped down. As X
3763 is now capable of enumerating graphics devices via udev in a
3764 seat-aware way the wrapper is not strictly necessary
3765 anymore. A stripped down temporary stop-gap is still shipped
3766 until the upstream display managers have been updated to
3767 fully support the new X logic. Expect this wrapper to be
3768 removed entirely in one of the next releases.
3770 * HandleSleepKey= in logind.conf has been split up into
3771 HandleSuspendKey= and HandleHibernateKey=. The old setting
3772 is not available anymore. X11 and the kernel are
3773 distuingishing between these keys and we should too. This
3774 also means the inhibition lock for these keys has been split
3777 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
3778 Poettering, Lukas Nykryn, Václav Pavlín
3782 * Whenever a unit changes state we will now log this to the
3783 journal and show along the unit's own log output in
3786 * ConditionPathIsMountPoint= can now properly detect bind
3787 mount points too. (Previously, a bind mount of one file
3788 system to another place in the same file system could not be
3789 detected as mount, since they shared struct stat's st_dev
3792 * We will now mount the cgroup controllers cpu, cpuacct,
3793 cpuset and the controllers net_cls, net_prio together by
3796 * nspawn containers will now have a virtualized boot
3797 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
3798 over with a randomized ID at container initialization). This
3799 has the effect of making "journalctl -b" do the right thing
3802 * The JSON output journal serialization has been updated not
3803 to generate "endless" list objects anymore, but rather one
3804 JSON object per line. This is more in line how most JSON
3805 parsers expect JSON objects. The new output mode
3806 "json-pretty" has been added to provide similar output, but
3807 neatly aligned for readability by humans.
3809 * We dropped all explicit sync() invocations in the shutdown
3810 code. The kernel does this implicitly anyway in the kernel
3811 reboot() syscall. halt(8)'s -n option is now a compatibility
3814 * We now support virtualized reboot() in containers, as
3815 supported by newer kernels. We will fall back to exit() if
3816 CAP_SYS_REBOOT is not available to the container. Also,
3817 nspawn makes use of this now and will actually reboot the
3818 container if the containerized OS asks for that.
3820 * journalctl will only show local log output by default
3821 now. Use --merge (-m) to show remote log output, too.
3823 * libsystemd-journal gained the new sd_journal_get_usage()
3824 call to determine the current disk usage of all journal
3825 files. This is exposed in the new "journalctl --disk-usage"
3828 * journald gained a new configuration setting SplitMode= in
3829 journald.conf which may be used to control how user journals
3830 are split off. See journald.conf(5) for details.
3832 * A new condition type ConditionFileNotEmpty= has been added.
3834 * tmpfiles' "w" lines now support file globbing, to write
3835 multiple files at once.
3837 * We added Python bindings for the journal submission
3838 APIs. More Python APIs for a number of selected APIs will
3839 likely follow. Note that we intend to add native bindings
3840 only for the Python language, as we consider it common
3841 enough to deserve bindings shipped within systemd. There are
3842 various projects outside of systemd that provide bindings
3843 for languages such as PHP or Lua.
3845 * Many conditions will now resolve specifiers such as %i. In
3846 addition, PathChanged= and related directives of .path units
3847 now support specifiers as well.
3849 * There's now a new RPM macro definition for the system preset
3852 * journald will now warn if it ca not forward a message to the
3853 syslog daemon because its socket is full.
3855 * timedated will no longer write or process /etc/timezone,
3856 except on Debian. As we do not support late mounted /usr
3857 anymore /etc/localtime always being a symlink is now safe,
3858 and hence the information in /etc/timezone is not necessary
3861 * logind will now always reserve one VT for a text getty (VT6
3862 by default). Previously if more than 6 X sessions where
3863 started they took up all the VTs with auto-spawned gettys,
3864 so that no text gettys were available anymore.
3866 * udev will now automatically inform the btrfs kernel logic
3867 about btrfs RAID components showing up. This should make
3868 simple hotplug based btrfs RAID assembly work.
3870 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
3871 (but not for its children which will stay at the kernel
3872 default). This should allow setups with a lot more listening
3875 * systemd will now always pass the configured timezone to the
3876 kernel at boot. timedated will do the same when the timezone
3879 * logind's inhibition logic has been updated. By default,
3880 logind will now handle the lid switch, the power and sleep
3881 keys all the time, even in graphical sessions. If DEs want
3882 to handle these events on their own they should take the new
3883 handle-power-key, handle-sleep-key and handle-lid-switch
3884 inhibitors during their runtime. A simple way to achieve
3885 that is to invoke the DE wrapped in an invocation of:
3887 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
3889 * Access to unit operations is now checked via SELinux taking
3890 the unit file label and client process label into account.
3892 * systemd will now notify the administrator in the journal
3893 when he over-mounts a non-empty directory.
3895 * There are new specifiers that are resolved in unit files,
3896 for the host name (%H), the machine ID (%m) and the boot ID
3899 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
3900 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
3901 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
3902 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
3903 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
3904 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
3905 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
3909 * Support for reading structured kernel messages from
3910 /dev/kmsg has now been added and is enabled by default.
3912 * Support for reading kernel messages from /proc/kmsg has now
3913 been removed. If you want kernel messages in the journal
3914 make sure to run a recent kernel (>= 3.5) that supports
3915 reading structured messages from /dev/kmsg (see
3916 above). /proc/kmsg is now exclusive property of classic
3917 syslog daemons again.
3919 * The libudev API gained the new
3920 udev_device_new_from_device_id() call.
3922 * The logic for file system namespace (ReadOnlyDirectory=,
3923 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
3924 require pivot_root() anymore. This means fewer temporary
3925 directories are created below /tmp for this feature.
3927 * nspawn containers will now see and receive all submounts
3928 made on the host OS below the root file system of the
3931 * Forward Secure Sealing is now supported for Journal files,
3932 which provide cryptographical sealing of journal files so
3933 that attackers cannot alter log history anymore without this
3934 being detectable. Lennart will soon post a blog story about
3935 this explaining it in more detail.
3937 * There are two new service settings RestartPreventExitStatus=
3938 and SuccessExitStatus= which allow configuration of exit
3939 status (exit code or signal) which will be excepted from the
3940 restart logic, resp. consider successful.
3942 * journalctl gained the new --verify switch that can be used
3943 to check the integrity of the structure of journal files and
3944 (if Forward Secure Sealing is enabled) the contents of
3947 * nspawn containers will now be run with /dev/stdin, /dev/fd/
3948 and similar symlinks pre-created. This makes running shells
3949 as container init process a lot more fun.
3951 * The fstab support can now handle PARTUUID= and PARTLABEL=
3954 * A new ConditionHost= condition has been added to match
3955 against the hostname (with globs) and machine ID. This is
3956 useful for clusters where a single OS image is used to
3957 provision a large number of hosts which shall run slightly
3958 different sets of services.
3960 * Services which hit the restart limit will now be placed in a
3963 Contributions from: Bertram Poettering, Dave Reisner, Huang
3964 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
3965 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
3969 * When running in --user mode systemd will now become a
3970 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
3971 tree a lot more organized.
3973 * A new PartOf= unit dependency type has been introduced that
3974 may be used to group services in a natural way.
3976 * "systemctl enable" may now be used to enable instances of
3979 * journalctl now prints error log levels in red, and
3980 warning/notice log levels in bright white. It also supports
3981 filtering by log level now.
3983 * cgtop gained a new -n switch (similar to top), to configure
3984 the maximum number of iterations to run for. It also gained
3985 -b, to run in batch mode (accepting no input).
3987 * The suffix ".service" may now be omitted on most systemctl
3988 command lines involving service unit names.
3990 * There's a new bus call in logind to lock all sessions, as
3991 well as a loginctl verb for it "lock-sessions".
3993 * libsystemd-logind.so gained a new call sd_journal_perror()
3994 that works similar to libc perror() but logs to the journal
3995 and encodes structured information about the error number.
3997 * /etc/crypttab entries now understand the new keyfile-size=
4000 * shutdown(8) now can send a (configurable) wall message when
4001 a shutdown is cancelled.
4003 * The mount propagation mode for the root file system will now
4004 default to "shared", which is useful to make containers work
4005 nicely out-of-the-box so that they receive new mounts from
4006 the host. This can be undone locally by running "mount
4007 --make-rprivate /" if needed.
4009 * The prefdm.service file has been removed. Distributions
4010 should maintain this unit downstream if they intend to keep
4011 it around. However, we recommend writing normal unit files
4012 for display managers instead.
4014 * Since systemd is a crucial part of the OS we will now
4015 default to a number of compiler switches that improve
4016 security (hardening) such as read-only relocations, stack
4017 protection, and suchlike.
4019 * The TimeoutSec= setting for services is now split into
4020 TimeoutStartSec= and TimeoutStopSec= to allow configuration
4021 of individual time outs for the start and the stop phase of
4024 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
4025 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
4026 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
4027 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
4028 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
4029 Gundersen, Zbigniew Jędrzejewski-Szmek
4033 * The journal and id128 C APIs are now fully documented as man
4036 * Extra safety checks have been added when transitioning from
4037 the initial RAM disk to the main system to avoid accidental
4040 * /etc/crypttab entries now understand the new keyfile-offset=
4043 * systemctl -t can now be used to filter by unit load state.
4045 * The journal C API gained the new sd_journal_wait() call to
4046 make writing synchronous journal clients easier.
4048 * journalctl gained the new -D switch to show journals from a
4051 * journalctl now displays a special marker between log
4052 messages of two different boots.
4054 * The journal is now explicitly flushed to /var via a service
4055 systemd-journal-flush.service, rather than implicitly simply
4056 by seeing /var/log/journal to be writable.
4058 * journalctl (and the journal C APIs) can now match for much
4059 more complex expressions, with alternatives and
4062 * When transitioning from the initial RAM disk to the main
4063 system we will now kill all processes in a killing spree to
4064 ensure no processes stay around by accident.
4066 * Three new specifiers may be used in unit files: %u, %h, %s
4067 resolve to the user name, user home directory resp. user
4068 shell. This is useful for running systemd user instances.
4070 * We now automatically rotate journal files if their data
4071 object hash table gets a fill level > 75%. We also size the
4072 hash table based on the configured maximum file size. This
4073 together should lower hash collisions drastically and thus
4074 speed things up a bit.
4076 * journalctl gained the new "--header" switch to introspect
4077 header data of journal files.
4079 * A new setting SystemCallFilters= has been added to services
4080 which may be used to apply blacklists or whitelists to
4081 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
4083 * nspawn gained a new --link-journal= switch (and quicker: -j)
4084 to link the container journal with the host. This makes it
4085 very easy to centralize log viewing on the host for all
4086 guests while still keeping the journal files separated.
4088 * Many bugfixes and optimizations
4090 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
4091 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
4092 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
4097 * Several tools now understand kernel command line arguments,
4098 which are only read when run in an initial RAM disk. They
4099 usually follow closely their normal counterparts, but are
4102 * There's a new tool to analyze the readahead files that are
4103 automatically generated at boot. Use:
4105 /usr/lib/systemd/systemd-readahead analyze /.readahead
4107 * We now provide an early debug shell on tty9 if this enabled. Use:
4109 systemctl enable debug-shell.service
4111 * All plymouth related units have been moved into the Plymouth
4112 package. Please make sure to upgrade your Plymouth version
4115 * systemd-tmpfiles now supports getting passed the basename of
4116 a configuration file only, in which case it will look for it
4117 in all appropriate directories automatically.
4119 * udevadm info now takes a /dev or /sys path as argument, and
4120 does the right thing. Example:
4122 udevadm info /dev/sda
4123 udevadm info /sys/class/block/sda
4125 * systemctl now prints a warning if a unit is stopped but a
4126 unit that might trigger it continues to run. Example: a
4127 service is stopped but the socket that activates it is left
4130 * "systemctl status" will now mention if the log output was
4131 shortened due to rotation since a service has been started.
4133 * The journal API now exposes functions to determine the
4134 "cutoff" times due to rotation.
4136 * journald now understands SIGUSR1 and SIGUSR2 for triggering
4137 immediately flushing of runtime logs to /var if possible,
4138 resp. for triggering immediate rotation of the journal
4141 * It is now considered an error if a service is attempted to
4142 be stopped that is not loaded.
4144 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
4146 * systemd-analyze now supports Python 3
4148 * tmpfiles now supports cleaning up directories via aging
4149 where the first level dirs are always kept around but
4150 directories beneath it automatically aged. This is enabled
4151 by prefixing the age field with '~'.
4153 * Seat objects now expose CanGraphical, CanTTY properties
4154 which is required to deal with very fast bootups where the
4155 display manager might be running before the graphics drivers
4156 completed initialization.
4158 * Seat objects now expose a State property.
4160 * We now include RPM macros for service enabling/disabling
4161 based on the preset logic. We recommend RPM based
4162 distributions to make use of these macros if possible. This
4163 makes it simpler to reuse RPM spec files across
4166 * We now make sure that the collected systemd unit name is
4167 always valid when services log to the journal via
4170 * There's a new man page kernel-command-line(7) detailing all
4171 command line options we understand.
4173 * The fstab generator may now be disabled at boot by passing
4174 fstab=0 on the kernel command line.
4176 * A new kernel command line option modules-load= is now understood
4177 to load a specific kernel module statically, early at boot.
4179 * Unit names specified on the systemctl command line are now
4180 automatically escaped as needed. Also, if file system or
4181 device paths are specified they are automatically turned
4182 into the appropriate mount or device unit names. Example:
4184 systemctl status /home
4185 systemctl status /dev/sda
4187 * The SysVConsole= configuration option has been removed from
4188 system.conf parsing.
4190 * The SysV search path is no longer exported on the D-Bus
4193 * The Names= option is been removed from unit file parsing.
4195 * There's a new man page bootup(7) detailing the boot process.
4197 * Every unit and every generator we ship with systemd now
4198 comes with full documentation. The self-explanatory boot is
4201 * A couple of services gained "systemd-" prefixes in their
4202 name if they wrap systemd code, rather than only external
4203 code. Among them fsck@.service which is now
4204 systemd-fsck@.service.
4206 * The HaveWatchdog property has been removed from the D-Bus
4209 * systemd.confirm_spawn= on the kernel command line should now
4212 * There's a new man page crypttab(5) which details all options
4213 we actually understand.
4215 * systemd-nspawn gained a new --capability= switch to pass
4216 additional capabilities to the container.
4218 * timedated will now read known NTP implementation unit names
4219 from /usr/lib/systemd/ntp-units.d/*.list,
4220 systemd-timedated-ntp.target has been removed.
4222 * journalctl gained a new switch "-b" that lists log data of
4223 the current boot only.
4225 * The notify socket is in the abstract namespace again, in
4226 order to support daemons which chroot() at start-up.
4228 * There is a new Storage= configuration option for journald
4229 which allows configuration of where log data should go. This
4230 also provides a way to disable journal logging entirely, so
4231 that data collected is only forwarded to the console, the
4232 kernel log buffer or another syslog implementation.
4234 * Many bugfixes and optimizations
4236 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
4237 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
4238 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
4239 Shawn Landden, Tom Gundersen
4243 * "systemctl help <unit>" now shows the man page if one is
4246 * Several new man pages have been added.
4248 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
4249 MaxLevelConsole= can now be specified in
4250 journald.conf. These options allow reducing the amount of
4251 data stored on disk or forwarded by the log level.
4253 * TimerSlackNSec= can now be specified in system.conf for
4254 PID1. This allows system-wide power savings.
4256 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
4257 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
4262 * logind is now capable of (optionally) handling power and
4263 sleep keys as well as the lid switch.
4265 * journalctl now understands the syntax "journalctl
4266 /usr/bin/avahi-daemon" to get all log output of a specific
4269 * CapabilityBoundingSet= in system.conf now also influences
4270 the capability bound set of usermode helpers of the kernel.
4272 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
4273 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
4274 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
4275 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
4279 * Note that we skipped 139 releases here in order to set the
4280 new version to something that is greater than both udev's
4281 and systemd's most recent version number.
4283 * udev: all udev sources are merged into the systemd source tree now.
4284 All future udev development will happen in the systemd tree. It
4285 is still fully supported to use the udev daemon and tools without
4286 systemd running, like in initramfs or other init systems. Building
4287 udev though, will require the *build* of the systemd tree, but
4288 udev can be properly *run* without systemd.
4290 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
4291 should be used to create dead device nodes as workarounds for broken
4294 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
4295 no longer supported. udev_monitor_new_from_netlink() needs to be
4296 used to subscribe to events.
4298 * udev: when udevd is started by systemd, processes which are left
4299 behind by forking them off of udev rules, are unconditionally cleaned
4300 up and killed now after the event handling has finished. Services or
4301 daemons must be started as systemd services. Services can be
4302 pulled-in by udev to get started, but they can no longer be directly
4303 forked by udev rules.
4305 * udev: the daemon binary is called systemd-udevd now and installed
4306 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
4307 to adapt to that, create symlink, or rename the binary after building
4310 * libudev no longer provides these symbols:
4311 udev_monitor_from_socket()
4312 udev_queue_get_failed_list_entry()
4313 udev_get_{dev,sys,run}_path()
4314 The versions number was bumped and symbol versioning introduced.
4316 * systemd-loginctl and systemd-journalctl have been renamed
4317 to loginctl and journalctl to match systemctl.
4319 * The config files: /etc/systemd/systemd-logind.conf and
4320 /etc/systemd/systemd-journald.conf have been renamed to
4321 logind.conf and journald.conf. Package updates should rename
4322 the files to the new names on upgrade.
4324 * For almost all files the license is now LGPL2.1+, changed
4325 from the previous GPL2.0+. Exceptions are some minor stuff
4326 of udev (which will be changed to LGPL2.1 eventually, too),
4327 and the MIT licensed sd-daemon.[ch] library that is suitable
4328 to be used as drop-in files.
4330 * systemd and logind now handle system sleep states, in
4331 particular suspending and hibernating.
4333 * logind now implements a sleep/shutdown/idle inhibiting logic
4334 suitable for a variety of uses. Soonishly Lennart will blog
4335 about this in more detail.
4337 * var-run.mount and var-lock.mount are no longer provided
4338 (which prevously bind mounted these directories to their new
4339 places). Distributions which have not converted these
4340 directories to symlinks should consider stealing these files
4341 from git history and add them downstream.
4343 * We introduced the Documentation= field for units and added
4344 this to all our shipped units. This is useful to make it
4345 easier to explore the boot and the purpose of the various
4348 * All smaller setup units (such as
4349 systemd-vconsole-setup.service) now detect properly if they
4350 are run in a container and are skipped when
4351 appropriate. This guarantees an entirely noise-free boot in
4352 Linux container environments such as systemd-nspawn.
4354 * A framework for implementing offline system updates is now
4355 integrated, for details see:
4356 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
4358 * A new service type Type=idle is available now which helps us
4359 avoiding ugly interleaving of getty output and boot status
4362 * There's now a system-wide CapabilityBoundingSet= option to
4363 globally reduce the set of capabilities for the
4364 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
4365 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
4366 even CAP_NET_ADMIN system-wide for secure systems.
4368 * There are now system-wide DefaultLimitXXX= options to
4369 globally change the defaults of the various resource limits
4370 for all units started by PID 1.
4372 * Harald Hoyer's systemd test suite has been integrated into
4373 systemd which allows easy testing of systemd builds in qemu
4374 and nspawn. (This is really awesome! Ask us for details!)
4376 * The fstab parser is now implemented as generator, not inside
4379 * systemctl will now warn you if .mount units generated from
4380 /etc/fstab are out of date due to changes in fstab that
4381 have not been read by systemd yet.
4383 * systemd is now suitable for usage in initrds. Dracut has
4384 already been updated to make use of this. With this in place
4385 initrds get a slight bit faster but primarily are much
4386 easier to introspect and debug since "systemctl status" in
4387 the host system can be used to introspect initrd services,
4388 and the journal from the initrd is kept around too.
4390 * systemd-delta has been added, a tool to explore differences
4391 between user/admin configuration and vendor defaults.
4393 * PrivateTmp= now affects both /tmp and /var/tmp.
4395 * Boot time status messages are now much prettier and feature
4396 proper english language. Booting up systemd has never been
4399 * Read-ahead pack files now include the inode number of all
4400 files to pre-cache. When the inode changes the pre-caching
4401 is not attempted. This should be nicer to deal with updated
4402 packages which might result in changes of read-ahead
4405 * We now temporaritly lower the kernel's read_ahead_kb variable
4406 when collecting read-ahead data to ensure the kernel's
4407 built-in read-ahead does not add noise to our measurements
4408 of necessary blocks to pre-cache.
4410 * There's now RequiresMountsFor= to add automatic dependencies
4411 for all mounts necessary for a specific file system path.
4413 * MountAuto= and SwapAuto= have been removed from
4414 system.conf. Mounting file systems at boot has to take place
4417 * nspawn now learned a new switch --uuid= to set the machine
4418 ID on the command line.
4420 * nspawn now learned the -b switch to automatically search
4423 * vt102 is now the default TERM for serial TTYs, upgraded from
4426 * systemd-logind now works on VT-less systems.
4428 * The build tree has been reorganized. The individual
4429 components now have directories of their own.
4431 * A new condition type ConditionPathIsReadWrite= is now available.
4433 * nspawn learned the new -C switch to create cgroups for the
4434 container in other hierarchies.
4436 * We now have support for hardware watchdogs, configurable in
4439 * The scheduled shutdown logic now has a public API.
4441 * We now mount /tmp as tmpfs by default, but this can be
4442 masked and /etc/fstab can override it.
4444 * Since udisks does not make use of /media anymore we are not
4445 mounting a tmpfs on it anymore.
4447 * journalctl gained a new --local switch to only interleave
4448 locally generated journal files.
4450 * We can now load the IMA policy at boot automatically.
4452 * The GTK tools have been split off into a systemd-ui.
4454 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
4455 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
4456 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
4457 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
4458 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
4459 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
4460 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
4461 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
4462 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
4467 * This is mostly a bugfix release
4469 * Support optional initialization of the machine ID from the
4470 KVM or container configured UUID.
4472 * Support immediate reboots with "systemctl reboot -ff"
4474 * Show /etc/os-release data in systemd-analyze output
4476 * Many bugfixes for the journal, including endianness fixes and
4477 ensuring that disk space enforcement works
4479 * sd-login.h is C++ comptaible again
4481 * Extend the /etc/os-release format on request of the Debian
4484 * We now refuse non-UTF8 strings used in various configuration
4485 and unit files. This is done to ensure we do not pass invalid
4486 data over D-Bus or expose it elsewhere.
4488 * Register Mimo USB Screens as suitable for automatic seat
4491 * Read SELinux client context from journal clients in a race
4494 * Reorder configuration file lookup order. /etc now always
4495 overrides /run in order to allow the administrator to always
4496 and unconditionally override vendor supplied or
4497 automatically generated data.
4499 * The various user visible bits of the journal now have man
4500 pages. We still lack man pages for the journal API calls
4503 * We now ship all man pages in HTML format again in the
4506 Contributions from: Dave Reisner, Dirk Eibach, Frederic
4507 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
4508 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
4513 * This is mostly a bugfix release
4515 * systems lacking /etc/os-release are no longer supported.
4517 * Various functionality updates to libsystemd-login.so
4519 * Track class of PAM logins to distuingish greeters from
4522 Contributions from: Kay Sievers, Lennart Poettering, Michael
4527 * This is an important bugfix release for v41.
4529 * Building man pages is now optional which should be useful
4530 for those building systemd from git but unwilling to install
4533 * Watchdog support for supervising services is now usable. In
4534 a future release support for hardware watchdogs
4535 (i.e. /dev/watchdog) will be added building on this.
4537 * Service start rate limiting is now configurable and can be
4538 turned off per service. When a start rate limit is hit a
4539 reboot can automatically be triggered.
4541 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
4543 Contributions from: Benjamin Franzke, Bill Nottingham,
4544 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
4545 Schmidt, Michał Górny, Piotr Drąg
4549 * The systemd binary is installed /usr/lib/systemd/systemd now;
4550 An existing /sbin/init symlink needs to be adapted with the
4553 * The code that loads kernel modules has been ported to invoke
4554 libkmod directly, instead of modprobe. This means we do not
4555 support systems with module-init-tools anymore.
4557 * Watchdog support is now already useful, but still not
4560 * A new kernel command line option systemd.setenv= is
4561 understood to set system wide environment variables
4562 dynamically at boot.
4564 * We now limit the set of capabilities of systemd-journald.
4566 * We now set SIGPIPE to ignore by default, since it only is
4567 useful in shell pipelines, and has little use in general
4568 code. This can be disabled with IgnoreSIPIPE=no in unit
4571 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
4572 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
4577 * This is mostly a bugfix release
4579 * We now expose the reason why a service failed in the
4580 "Result" D-Bus property.
4582 * Rudimentary service watchdog support (will be completed over
4583 the next few releases.)
4585 * When systemd forks off in order execute some service we will
4586 now immediately changes its argv[0] to reflect which process
4587 it will execute. This is useful to minimize the time window
4588 with a generic argv[0], which makes bootcharts more useful
4590 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
4591 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
4592 Mike Kazantsev, Ray Strode
4596 * This is mostly a test release, but incorporates many
4599 * New systemd-cgtop tool to show control groups by their
4602 * Linking against libacl for ACLs is optional again. If
4603 disabled, support tracking device access for active logins
4604 goes becomes unavailable, and so does access to the user
4605 journals by the respective users.
4607 * If a group "adm" exists, journal files are automatically
4608 owned by them, thus allow members of this group full access
4609 to the system journal as well as all user journals.
4611 * The journal now stores the SELinux context of the logging
4612 client for all entries.
4614 * Add C++ inclusion guards to all public headers
4616 * New output mode "cat" in the journal to print only text
4617 messages, without any meta data like date or time.
4619 * Include tiny X server wrapper as a temporary stop-gap to
4620 teach XOrg udev display enumeration. This is used by display
4621 managers such as gdm, and will go away as soon as XOrg
4622 learned native udev hotplugging for display devices.
4624 * Add new systemd-cat tool for executing arbitrary programs
4625 with STDERR/STDOUT connected to the journal. Can also act as
4626 BSD logger replacement, and does so by default.
4628 * Optionally store all locally generated coredumps in the
4629 journal along with meta data.
4631 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
4632 writing short strings to files (for usage for /sys), and for
4633 creating symlinks, character and block device nodes.
4635 * New unit file option ControlGroupPersistent= to make cgroups
4636 persistent, following the mechanisms outlined in
4637 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
4639 * Support multiple local RTCs in a sane way
4641 * No longer monopolize IO when replaying readahead data on
4642 rotating disks, since we might starve non-file-system IO to
4643 death, since fanotify() will not see accesses done by blkid,
4646 * Do not show kernel threads in systemd-cgls anymore, unless
4647 requested with new -k switch.
4649 Contributions from: Dan Horák, Kay Sievers, Lennart
4650 Poettering, Michal Schmidt
4654 * This is mostly a test release, but incorporates many
4657 * The git repository moved to:
4658 git://anongit.freedesktop.org/systemd/systemd
4659 ssh://git.freedesktop.org/git/systemd/systemd
4661 * First release with the journal
4662 http://0pointer.de/blog/projects/the-journal.html
4664 * The journal replaces both systemd-kmsg-syslogd and
4665 systemd-stdout-bridge.
4667 * New sd_pid_get_unit() API call in libsystemd-logind
4669 * Many systemadm clean-ups
4671 * Introduce remote-fs-pre.target which is ordered before all
4672 remote mounts and may be used to start services before all
4675 * Added Mageia support
4677 * Add bash completion for systemd-loginctl
4679 * Actively monitor PID file creation for daemons which exit in
4680 the parent process before having finished writing the PID
4681 file in the daemon process. Daemons which do this need to be
4682 fixed (i.e. PID file creation must have finished before the
4683 parent exits), but we now react a bit more gracefully to them.
4685 * Add colourful boot output, mimicking the well-known output
4686 of existing distributions.
4688 * New option PassCredentials= for socket units, for
4689 compatibility with a recent kernel ABI breakage.
4691 * /etc/rc.local is now hooked in via a generator binary, and
4692 thus will no longer act as synchronization point during
4695 * systemctl list-unit-files now supports --root=.
4697 * systemd-tmpfiles now understands two new commands: z, Z for
4698 relabelling files according to the SELinux database. This is
4699 useful to apply SELinux labels to specific files in /sys,
4702 * Output of SysV services is now forwarded to both the console
4703 and the journal by default, not only just the console.
4705 * New man pages for all APIs from libsystemd-login.
4707 * The build tree got reorganized and a the build system is a
4708 lot more modular allowing embedded setups to specifically
4709 select the components of systemd they are interested in.
4711 * Support for Linux systems lacking the kernel VT subsystem is
4714 * configure's --with-rootdir= got renamed to
4715 --with-rootprefix= to follow the naming used by udev and
4718 * Unless specified otherwise we will now install to /usr instead
4719 of /usr/local by default.
4721 * Processes with '@' in argv[0][0] are now excluded from the
4722 final shut-down killing spree, following the logic explained
4724 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
4726 * All processes remaining in a service cgroup when we enter
4727 the START or START_PRE states are now killed with
4728 SIGKILL. That means it is no longer possible to spawn
4729 background processes from ExecStart= lines (which was never
4730 supported anyway, and bad style).
4732 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
4733 reloading of units together.
4735 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
4736 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
4737 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
4738 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
4739 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek