1 systemd System and Service Manager
5 * When querying unit file enablement status (for example via
6 "systemctl is-enabled"), a new state "indirect" is now known
7 which indicates that a unit might not be enabled itself, but
8 another unit listed in its Alias= setting might be.
10 * Similar to the various existing ConditionXYZ= settings for
11 units there are now matching AssertXYZ= settings. While
12 failing conditions cause a unit to be skipped, but its job
13 to succeed, failing assertions declared like this will cause
14 a unit start operation and its job to fail.
16 * hostnamed now knows a new chassis type "embedded".
18 * systemctl gained a new "edit" command. When used on a unit
19 file this allows extending unit files with .d/ drop-in
20 configuration snippets or editing the full file (after
21 copying it from /usr/lib to /etc). This will invoke the
22 user's editor (as configured with $EDITOR), and reload the
23 modified configuration after editing.
25 * "systemctl status" now shows the suggested enablement state
26 for a unit, as declared in the (usually vendor-supplied)
29 * nss-myhostname will now resolve the single-label host name
30 "gateway" to the locally configured default IP routing
31 gateways, ordered by their metrics. This assigns a stable
32 name to the used gateways, regardless which ones are
33 currently configured. Note that the name will only be
34 resolved after all other name sources (if nss-myhostname is
35 configured properly) and should hence not negatively impact
36 systems that use the single-label host name "gateway" in
39 * systemd-inhibit now allows filtering by mode when listing
42 * Scope and service units gained a new "Delegate" boolean
43 property, which when set allows processes running inside the
44 unit to further partition resources. This is primarily
45 useful for systemd user instances as well as container
48 * journald will now pick up audit messages directly from
49 the kernel, and log them like any other log message. The
50 audit fields are split up and fully indexed. This means that
51 journalctl in many ways is now a (nicer!) alternative to
52 ausearch, the traditional audit client. Note that this
53 implements only a minimal audit client, if you want the
54 special audit modes like reboot-on-log-overflow, please use
55 the traditional auditd instead, which can be used in
58 * The ConditionSecurity= unit file option now understands the
59 special string "audit" to check whether auditing is
62 * journalctl gained two new commands --vacuum-size= and
63 --vacuum-time= to delete old journal files until the
64 remaining ones take up no more the specified size on disk,
65 or are not older than the specified time.
67 * A new, native PPPoE library has been added to sd-network,
68 systemd's library of light-weight networking protocols. This
69 library will be used in a future version of networkd to
70 enable PPPoE communication without an external pppd daemon.
72 * The busctl tool now understands a new "capture" verb that
73 works similar to "monitor", but writes a packet capture
74 trace to STDOUT that can be redirected to a file which is
75 compatible with libcap's capture file format. This can then
76 be loaded in Wireshark and similar tools to inspect bus
79 * The busctl tool now understands a new "tree" verb that shows
80 the object trees of a specific service on the bus, or of all
83 * The busctl tool now understands a new "introspect" verb that
84 shows all interfaces and members of objects on the bus,
85 including their signature and values. This is particularly
86 useful to get more information about bus objects shown by
87 the new "busctl tree" command.
89 * The busctl tool now understands new verbs "call",
90 "set-property" and "get-property" for invoking bus method
91 calls, setting and getting bus object properties in a
94 * busctl gained a new --augment-creds= argument that controls
95 whether the tool shall augment credential information it
96 gets from the bus with data from /proc, in a possibly
99 * nspawn's --link-journal= switch gained two new values
100 "try-guest" and "try-host" that work like "guest" and
101 "host", but do not fail if the host has no persistent
102 journalling enabled. -j is now equivalent to
103 --link-journal=try-guest.
105 * macvlan network devices created by nspawn will now have
106 stable MAC addresses.
108 * A new SmackProcessLabel= unit setting has been added, which
109 controls the SMACK security label processes forked off by
110 the respective unit shall use.
112 * If compiled with --enable-xkbcommon, systemd-localed will
113 verify x11 keymap settings by compiling the given keymap. It
114 will spew out warnings if the compilation fails. This
115 requires libxkbcommon to be installed.
117 * When a coredump is collected a larger number of metadata
118 fields is now collected and included in the journal records
119 created for it. More specifically control group membership,
120 environment variables, memory maps, working directory,
121 chroot directory, /proc/$PID/status, and a list of open file
122 descriptors is now stored in the log entry.
124 * The udev hwdb now contains DPI information for mice. For
127 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
129 * All systemd programs that read standalone configuration
130 files in /etc now also support a corresponding series of
131 .conf.d configuration directories in /etc/, /run/,
132 /usr/local/lib/, /usr/lib/, and (if configured with
133 --enable-split-usr) /lib/. In particular, the following
134 configuration files now have corresponding configuration
135 directories: system.conf user.conf, logind.conf,
136 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
137 resolved.conf, timesyncd.conf, journal-remote.conf, and
138 journal-upload.conf. Note that distributions should use the
139 configuration directories in /usr/lib/; the directories in
140 /etc/ are reserved for the system administrator.
142 * systemd-rfkill will no longer take the rfkill device name
143 into account when storing rfkill state on disk, as the name
144 might be dynamically assigned and not stable. Instead, the
145 ID_PATH udev variable combined with the rfkill type (wlan,
146 bluetooth, ...) is used.
148 * A new service systemd-machine-id-commit.service has been
149 added. When used on systems where /etc is read-only during
150 boot, and /etc/machine-id is not initialized (but an empty
151 file), this service will copy the temporary machine ID
152 created as replacement into /etc after the system is fully
153 booted up. This is useful for systems that are freshly
154 installed with a non-initialized machine ID, but should get
155 a fixed machine ID for subsequent boots.
157 * networkd's .netdev files now provide a large set of
158 configuration parameters for VXLAN devices. Similar, the
159 bridge port cost parameter is now configurable in .network
160 files. There's also new support for configuring IP source
161 routing. networkd .link files gained support for a new
162 OriginalName= match that is useful to match against the
163 original interface name the kernel assigned. .network files
164 may include MTU= and MACAddress= fields for altering the MTU
165 and MAC address while being connected to a specific network
168 * The LUKS logic gained supported for configuring
169 UUID-specific key files. There's also new support for naming
170 LUKS device from the kernel command line, using the new
173 * Timer units may now be transiently created via the bus API
174 (this was previously already available for scope and service
175 units). In addition it is now possible to create multiple
176 transient units at the same time with a single bus call. The
177 "systemd-run" tool has been updated to make use of this for
178 running commands on a specified time, in at(1)-style.
180 * tmpfiles gained support for "t" lines, for assigning
181 extended attributes to files. Among other uses this may be
182 used to assign SMACK labels to files.
184 Contributions from: Alin Rauta, Alison Chaiken, Andrej
185 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
186 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
187 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
188 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
189 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
190 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
191 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
192 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
193 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
194 Michael Biebl, Michael Chapman, Michael Marineau, Michal
195 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
196 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
197 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
198 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
199 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
200 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
201 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
203 -- Berlin, 2014-12-10
207 * journalctl gained the new options -t/--identifier= to match
208 on the syslog identifier (aka "tag"), as well as --utc to
209 show log timestamps in the UTC timezone. journalctl now also
210 accepts -n/--lines=all to disable line capping in a pager.
212 * journalctl gained a new switch, --flush, that synchronously
213 flushes logs from /run/log/journal to /var/log/journal if
214 persistent storage is enabled. systemd-journal-flush.service
215 now waits until the operation is complete.
217 * Services can notify the manager before they start a reload
218 (by sending RELOADING=1) or shutdown (by sending
219 STOPPING=1). This allows the manager to track and show the
220 internal state of daemons and closes a race condition when
221 the process is still running but has closed its D-Bus
224 * Services with Type=oneshot do not have to have any ExecStart
227 * User units are now loaded also from
228 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
229 /run/systemd/user directory that was already previously
230 supported, but is under the control of the user.
232 * Job timeouts (i.e. time-outs on the time a job that is
233 queued stays in the run queue) can now optionally result in
234 immediate reboot or power-off actions (JobTimeoutAction= and
235 JobTimeoutRebootArgument=). This is useful on ".target"
236 units, to limit the maximum time a target remains
237 undispatched in the run queue, and to trigger an emergency
238 operation in such a case. This is now used by default to
239 turn off the system if boot-up (as defined by everything in
240 basic.target) hangs and does not complete for at least
241 15min. Also, if power-off or reboot hang for at least 30min
242 an immediate power-off/reboot operation is triggered. This
243 functionality is particularly useful to increase reliability
244 on embedded devices, but also on laptops which might
245 accidentally get powered on when carried in a backpack and
246 whose boot stays stuck in a hard disk encryption passphrase
249 * systemd-logind can be configured to also handle lid switch
250 events even when the machine is docked or multiple displays
251 are attached (HandleLidSwitchDocked= option).
253 * A helper binary and a service have been added which can be
254 used to resume from hibernation in the initramfs. A
255 generator will parse the resume= option on the kernel
256 command line to trigger resume.
258 * A user console daemon systemd-consoled has been
259 added. Currently, it is a preview, and will so far open a
260 single terminal on each session of the user marked as
261 Desktop=systemd-console.
263 * Route metrics can be specified for DHCP routes added by
266 * The SELinux context of socket-activated services can be set
267 from the information provided by the networking stack
268 (SELinuxContextFromNet= option).
270 * Userspace firmware loading support has been removed and
271 the minimum supported kernel version is thus bumped to 3.7.
273 * Timeout for udev workers has been increased from 1 to 3
274 minutes, but a warning will be printed after 1 minute to
275 help diagnose kernel modules that take a long time to load.
277 * Udev rules can now remove tags on devices with TAG-="foobar".
279 * systemd's readahead implementation has been removed. In many
280 circumstances it didn't give expected benefits even for
281 rotational disk drives and was becoming less relevant in the
282 age of SSDs. As none of the developers has been using
283 rotating media anymore, and nobody stepped up to actively
284 maintain this component of systemd it has now been removed.
286 * Swap units can use Options= to specify discard options.
287 Discard options specified for swaps in /etc/fstab are now
290 * Docker containers are now detected as a separate type of
293 * The Password Agent protocol gained support for queries where
294 the user input is shown, useful e.g. for user names.
295 systemd-ask-password gained a new --echo option to turn that
298 * The default sysctl.d/ snippets will now set:
300 net.core.default_qdisc = fq_codel
302 This selects Fair Queuing Controlled Delay as the default
303 queuing discipline for network interfaces. fq_codel helps
304 fight the network bufferbloat problem. It is believed to be
305 a good default with no tuning required for most workloads.
306 Downstream distributions may override this choice. On 10Gbit
307 servers that do not do forwarding, "fq" may perform better.
308 Systems without a good clocksource should use "pfifo_fast".
310 * If kdbus is enabled during build a new option BusPolicy= is
311 available for service units, that allows locking all service
312 processes into a stricter bus policy, in order to limit
313 access to various bus services, or even hide most of them
314 from the service's view entirely.
316 * networkctl will now show the .network and .link file
317 networkd has applied to a specific interface.
319 * sd-login gained a new API call sd_session_get_desktop() to
320 query which desktop environment has been selected for a
323 * UNIX utmp support is now compile-time optional to support
326 * systemctl gained two new commands "add-wants" and
327 "add-requires" for pulling in units from specific targets
330 * If the word "rescue" is specified on the kernel command line
331 the system will now boot into rescue mode (aka
332 rescue.target), which was previously available only by
333 specifying "1" or "systemd.unit=rescue.target" on the kernel
334 command line. This new kernel command line option nicely
335 mirrors the already existing "emergency" kernel command line
338 * New kernel command line options mount.usr=, mount.usrflags=,
339 mount.usrfstype= have been added that match root=, rootflags=,
340 rootfstype= but allow mounting a specific file system to
343 * The $NOTIFY_SOCKET is now also passed to control processes of
344 services, not only the main process.
346 * This version reenables support for fsck's -l switch. This
347 means at least version v2.25 of util-linux is required for
348 operation, otherwise dead-locks on device nodes may
349 occur. Again: you need to update util-linux to at least
350 v2.25 when updating systemd to v217.
352 * The "multi-seat-x" tool has been removed from systemd, as
353 its functionality has been integrated into X servers 1.16,
354 and the tool is hence redundant. It is recommended to update
355 display managers invoking this tool to simply invoke X
356 directly from now on, again.
358 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
359 message flag has been added for all of systemd's PolicyKit
360 authenticated method calls has been added. In particular
361 this now allows optional interactive authorization via
362 PolicyKit for many of PID1's privileged operations such as
363 unit file enabling and disabling.
365 * "udevadm hwdb --update" learnt a new switch "--usr" for
366 placing the rebuilt hardware database in /usr instead of
367 /etc. When used only hardware database entries stored in
368 /usr will be used, and any user database entries in /etc are
369 ignored. This functionality is useful for vendors to ship a
370 pre-built database on systems where local configuration is
371 unnecessary or unlikely.
373 * Calendar time specifications in .timer units now also
374 understand the strings "semi-annually", "quarterly" and
375 "minutely" as shortcuts (in addition to the preexisting
376 "anually", "hourly", ...).
378 * systemd-tmpfiles will now correctly create files in /dev
379 at boot which are marked for creation only at boot. It is
380 recommended to always create static device nodes with 'c!'
381 and 'b!', so that they are created only at boot and not
382 overwritten at runtime.
384 * When the watchdog logic is used for a service (WatchdogSec=)
385 and the watchdog timeout is hit the service will now be
386 terminated with SIGABRT (instead of just SIGTERM), in order
387 to make sure a proper coredump and backtrace is
388 generated. This ensures that hanging services will result in
389 similar coredump/backtrace behaviour as services that hit a
392 Contributions from: Andreas Henriksson, Andrei Borzenkov,
393 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
394 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
395 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
396 Herrmann, David Sommerseth, David Strauss, Emil Renner
397 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
398 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
399 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
400 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
401 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
402 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
403 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
404 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
405 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
406 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
407 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
408 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
409 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
410 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
411 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
412 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
415 -- Berlin, 2014-10-28
419 * timedated no longer reads NTP implementation unit names from
420 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
421 implementations should add a
423 Conflicts=systemd-timesyncd.service
425 to their unit files to take over and replace systemd's NTP
426 default functionality.
428 * systemd-sysusers gained a new line type "r" for configuring
429 which UID/GID ranges to allocate system users/groups
430 from. Lines of type "u" may now add an additional column
431 that specifies the home directory for the system user to be
432 created. Also, systemd-sysusers may now optionally read user
433 information from STDIN instead of a file. This is useful for
434 invoking it from RPM preinst scriptlets that need to create
435 users before the first RPM file is installed since these
436 files might need to be owned by them. A new
437 %sysusers_create_inline RPM macro has been introduced to do
438 just that. systemd-sysusers now updates the shadow files as
439 well as the user/group databases, which should enhance
440 compatibility with certain tools like grpck.
442 * A number of bus APIs of PID 1 now optionally consult
443 PolicyKit to permit access for otherwise unprivileged
444 clients under certain conditions. Note that this currently
445 doesn't support interactive authentication yet, but this is
446 expected to be added eventually, too.
448 * /etc/machine-info now has new fields for configuring the
449 deployment environment of the machine, as well as the
450 location of the machine. hostnamectl has been updated with
451 new command to update these fields.
453 * systemd-timesyncd has been updated to automatically acquire
454 NTP server information from systemd-networkd, which might
455 have been discovered via DHCP.
457 * systemd-resolved now includes a caching DNS stub resolver
458 and a complete LLMNR name resolution implementation. A new
459 NSS module "nss-resolve" has been added which can be used
460 instead of glibc's own "nss-dns" to resolve hostnames via
461 systemd-resolved. Hostnames, addresses and arbitrary RRs may
462 be resolved via systemd-resolved D-Bus APIs. In contrast to
463 the glibc internal resolver systemd-resolved is aware of
464 multi-homed system, and keeps DNS server and caches separate
465 and per-interface. Queries are sent simultaneously on all
466 interfaces that have DNS servers configured, in order to
467 properly handle VPNs and local LANs which might resolve
468 separate sets of domain names. systemd-resolved may acquire
469 DNS server information from systemd-networkd automatically,
470 which in turn might have discovered them via DHCP. A tool
471 "systemd-resolve-host" has been added that may be used to
472 query the DNS logic in resolved. systemd-resolved implements
473 IDNA and automatically uses IDNA or UTF-8 encoding depending
474 on whether classic DNS or LLMNR is used as transport. In the
475 next releases we intend to add a DNSSEC and mDNS/DNS-SD
476 implementation to systemd-resolved.
478 * A new NSS module nss-mymachines has been added, that
479 automatically resolves the names of all local registered
480 containers to their respective IP addresses.
482 * A new client tool "networkctl" for systemd-networkd has been
483 added. It currently is entirely passive and will query
484 networking configuration from udev, rtnetlink and networkd,
485 and present it to the user in a very friendly
486 way. Eventually, we hope to extend it to become a full
487 control utility for networkd.
489 * .socket units gained a new DeferAcceptSec= setting that
490 controls the kernels' TCP_DEFER_ACCEPT sockopt for
491 TCP. Similar, support for controlling TCP keep-alive
492 settings has been added (KeepAliveTimeSec=,
493 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
494 turning off Nagle's algorithm on TCP has been added
497 * logind learned a new session type "web", for use in projects
498 like Cockpit which register web clients as PAM sessions.
500 * timer units with at least one OnCalendar= setting will now
501 be started only after timer-sync.target has been
502 reached. This way they will not elapse before the system
503 clock has been corrected by a local NTP client or
504 similar. This is particular useful on RTC-less embedded
505 machines, that come up with an invalid system clock.
507 * systemd-nspawn's --network-veth= switch should now result in
508 stable MAC addresses for both the outer and the inner side
511 * systemd-nspawn gained a new --volatile= switch for running
512 container instances with /etc or /var unpopulated.
514 * The kdbus client code has been updated to use the new Linux
515 3.17 memfd subsystem instead of the old kdbus-specific one.
517 * systemd-networkd's DHCP client and server now support
518 FORCERENEW. There are also new configuration options to
519 configure the vendor client identifier and broadcast mode
522 * systemd will no longer inform the kernel about the current
523 timezone, as this is necessarily incorrect and racy as the
524 kernel has no understanding of DST and similar
525 concepts. This hence means FAT timestamps will be always
526 considered UTC, similar to what Android is already
527 doing. Also, when the RTC is configured to the local time
528 (rather than UTC) systemd will never synchronize back to it,
529 as this might confuse Windows at a later boot.
531 * systemd-analyze gained a new command "verify" for offline
532 validation of unit files.
534 * systemd-networkd gained support for a couple of additional
535 settings for bonding networking setups. Also, the metric for
536 statically configured routes may now be configured. For
537 network interfaces where this is appropriate the peer IP
538 address may now be configured.
540 * systemd-networkd's DHCP client will no longer request
541 broadcasting by default, as this tripped up some networks.
542 For hardware where broadcast is required the feature should
543 be switched back on using RequestBroadcast=yes.
545 * systemd-networkd will now set up IPv4LL addresses (when
546 enabled) even if DHCP is configured successfully.
548 * udev will now default to respect network device names given
549 by the kernel when the kernel indicates that these are
550 predictable. This behavior can be tweaked by changing
551 NamePolicy= in the relevant .link file.
553 * A new library systemd-terminal has been added that
554 implements full TTY stream parsing and rendering. This
555 library is supposed to be used later on for implementing a
556 full userspace VT subsystem, replacing the current kernel
559 * A new tool systemd-journal-upload has been added to push
560 journal data to a remote system running
561 systemd-journal-remote.
563 * journald will no longer forward all local data to another
564 running syslog daemon. This change has been made because
565 rsyslog (which appears to be the most commonly used syslog
566 implementation these days) no longer makes use of this, and
567 instead pulls the data out of the journal on its own. Since
568 forwarding the messages to a non-existent syslog server is
569 more expensive than we assumed we have now turned this
570 off. If you run a syslog server that is not a recent rsyslog
571 version, you have to turn this option on again
572 (ForwardToSyslog= in journald.conf).
574 * journald now optionally supports the LZ4 compressor for
575 larger journal fields. This compressor should perform much
576 better than XZ which was the previous default.
578 * machinectl now shows the IP addresses of local containers,
579 if it knows them, plus the interface name of the container.
581 * A new tool "systemd-escape" has been added that makes it
582 easy to escape strings to build unit names and similar.
584 * sd_notify() messages may now include a new ERRNO= field
585 which is parsed and collected by systemd and shown among the
586 "systemctl status" output for a service.
588 * A new component "systemd-firstboot" has been added that
589 queries the most basic systemd information (timezone,
590 hostname, root password) interactively on first
591 boot. Alternatively it may also be used to provision these
592 things offline on OS images installed into directories.
594 * The default sysctl.d/ snippets will now set
596 net.ipv4.conf.default.promote_secondaries=1
598 This has the benefit of no flushing secondary IP addresses
599 when primary addresses are removed.
601 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
602 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
603 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
604 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
605 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
606 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
607 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
608 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
609 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
610 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
611 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
612 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
613 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
614 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
615 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
617 -- Berlin, 2014-08-19
621 * A new tool systemd-sysusers has been added. This tool
622 creates system users and groups in /etc/passwd and
623 /etc/group, based on static declarative system user/group
624 definitions in /usr/lib/sysusers.d/. This is useful to
625 enable factory resets and volatile systems that boot up with
626 an empty /etc directory, and thus need system users and
627 groups created during early boot. systemd now also ships
628 with two default sysusers.d/ files for the most basic
629 users and groups systemd and the core operating system
632 * A new tmpfiles snippet has been added that rebuilds the
633 essential files in /etc on boot, should they be missing.
635 * A directive for ensuring automatic clean-up of
636 /var/cache/man/ has been removed from the default
637 configuration. This line should now be shipped by the man
638 implementation. The necessary change has been made to the
639 man-db implementation. Note that you need to update your man
640 implementation to one that ships this line, otherwise no
641 automatic clean-up of /var/cache/man will take place.
643 * A new condition ConditionNeedsUpdate= has been added that
644 may conditionalize services to only run when /etc or /var
645 are "older" than the vendor operating system resources in
646 /usr. This is useful for reconstructing or updating /etc
647 after an offline update of /usr or a factory reset, on the
648 next reboot. Services that want to run once after such an
649 update or reset should use this condition and order
650 themselves before the new systemd-update-done.service, which
651 will mark the two directories as fully updated. A number of
652 service files have been added making use of this, to rebuild
653 the udev hardware database, the journald message catalog and
654 dynamic loader cache (ldconfig). The systemd-sysusers tool
655 described above also makes use of this now. With this in
656 place it is now possible to start up a minimal operating
657 system with /etc empty cleanly. For more information on the
658 concepts involved see this recent blog story:
660 http://0pointer.de/blog/projects/stateless.html
662 * A new system group "input" has been introduced, and all
663 input device nodes get this group assigned. This is useful
664 for system-level software to get access to input devices. It
665 complements what is already done for "audio" and "video".
667 * systemd-networkd learnt minimal DHCPv4 server support in
668 addition to the existing DHCPv4 client support. It also
669 learnt DHCPv6 client and IPv6 Router Solicitation client
670 support. The DHCPv4 client gained support for static routes
671 passed in from the server. Note that the [DHCPv4] section
672 known in older systemd-networkd versions has been renamed to
673 [DHCP] and is now also used by the DHCPv6 client. Existing
674 .network files using settings of this section should be
675 updated, though compatibility is maintained. Optionally, the
676 client hostname may now be sent to the DHCP server.
678 * networkd gained support for vxlan virtual networks as well
679 as tun/tap and dummy devices.
681 * networkd gained support for automatic allocation of address
682 ranges for interfaces from a system-wide pool of
683 addresses. This is useful for dynamically managing a large
684 number of interfaces with a single network configuration
685 file. In particular this is useful to easily assign
686 appropriate IP addresses to the veth links of a large number
689 * RPM macros for processing sysusers, sysctl and binfmt
690 drop-in snippets at package installation time have been
693 * The /etc/os-release file should now be placed in
694 /usr/lib/os-release. The old location is automatically
695 created as symlink. /usr/lib is the more appropriate
696 location of this file, since it shall actually describe the
697 vendor operating system shipped in /usr, and not the
698 configuration stored in /etc.
700 * .mount units gained a new boolean SloppyOptions= setting
701 that maps to mount(8)'s -s option which enables permissive
702 parsing of unknown mount options.
704 * tmpfiles learnt a new "L+" directive which creates a symlink
705 but (unlike "L") deletes a pre-existing file first, should
706 it already exist and not already be the correct
707 symlink. Similar, "b+", "c+" and "p+" directives have been
708 added as well, which create block and character devices, as
709 well as fifos in the filesystem, possibly removing any
710 pre-existing files of different types.
712 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
713 'argument' field (which so far specified the source to
714 symlink/copy the files from) is now optional. If omitted the
715 same file os copied from /usr/share/factory/ suffixed by the
716 full destination path. This is useful for populating /etc
717 with essential files, by copying them from vendor defaults
718 shipped in /usr/share/factory/etc.
720 * A new command "systemctl preset-all" has been added that
721 applies the service preset settings to all installed unit
722 files. A new switch --preset-mode= has been added that
723 controls whether only enable or only disable operations
726 * A new command "systemctl is-system-running" has been added
727 that allows checking the overall state of the system, for
728 example whether it is fully up and running.
730 * When the system boots up with an empty /etc, the equivalent
731 to "systemctl preset-all" is executed during early boot, to
732 make sure all default services are enabled after a factory
735 * systemd now contains a minimal preset file that enables the
736 most basic services systemd ships by default.
738 * Unit files' [Install] section gained a new DefaultInstance=
739 field for defining the default instance to create if a
740 template unit is enabled with no instance specified.
742 * A new passive target cryptsetup-pre.target has been added
743 that may be used by services that need to make they run and
744 finish before the first LUKS cryptographic device is set up.
746 * The /dev/loop-control and /dev/btrfs-control device nodes
747 are now owned by the "disk" group by default, opening up
748 access to this group.
750 * systemd-coredump will now automatically generate a
751 stack trace of all core dumps taking place on the system,
752 based on elfutils' libdw library. This stack trace is logged
755 * systemd-coredump may now optionally store coredumps directly
756 on disk (in /var/lib/systemd/coredump, possibly compressed),
757 instead of storing them unconditionally in the journal. This
758 mode is the new default. A new configuration file
759 /etc/systemd/coredump.conf has been added to configure this
760 and other parameters of systemd-coredump.
762 * coredumpctl gained a new "info" verb to show details about a
763 specific coredump. A new switch "-1" has also been added
764 that makes sure to only show information about the most
765 recent entry instead of all entries. Also, as the tool is
766 generally useful now the "systemd-" prefix of the binary
767 name has been removed. Distributions that want to maintain
768 compatibility with the old name should add a symlink from
769 the old name to the new name.
771 * journald's SplitMode= now defaults to "uid". This makes sure
772 that unprivileged users can access their own coredumps with
773 coredumpctl without restrictions.
775 * New kernel command line options "systemd.wants=" (for
776 pulling an additional unit during boot), "systemd.mask="
777 (for masking a specific unit for the boot), and
778 "systemd.debug-shell" (for enabling the debug shell on tty9)
779 have been added. This is implemented in the new generator
780 "systemd-debug-generator".
782 * systemd-nspawn will now by default filter a couple of
783 syscalls for containers, among them those required for
784 kernel module loading, direct x86 IO port access, swap
785 management, and kexec. Most importantly though
786 open_by_handle_at() is now prohibited for containers,
787 closing a hole similar to a recently discussed vulnerability
788 in docker regarding access to files on file hierarchies the
789 container should normally not have access to. Note that for
790 nspawn we generally make no security claims anyway (and
791 this is explicitly documented in the man page), so this is
792 just a fix for one of the most obvious problems.
794 * A new man page file-hierarchy(7) has been added that
795 contains a minimized, modernized version of the file system
796 layout systemd expects, similar in style to the FHS
797 specification or hier(5). A new tool systemd-path(1) has
798 been added to query many of these paths for the local
801 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
802 longer done. Since the directory now has a per-user size
803 limit, and is cleaned on logout this appears unnecessary,
804 in particular since this now brings the lifecycle of this
805 directory closer in line with how IPC objects are handled.
807 * systemd.pc now exports a number of additional directories,
808 including $libdir (which is useful to identify the library
809 path for the primary architecture of the system), and a
810 couple of drop-in directories.
812 * udev's predictable network interface names now use the dev_port
813 sysfs attribute, introduced in linux 3.15 instead of dev_id to
814 distinguish between ports of the same PCI function. dev_id should
815 only be used for ports using the same HW address, hence the need
818 * machined has been updated to export the OS version of a
819 container (read from /etc/os-release and
820 /usr/lib/os-release) on the bus. This is now shown in
821 "machinectl status" for a machine.
823 * A new service setting RestartForceExitStatus= has been
824 added. If configured to a set of exit signals or process
825 return values, the service will be restarted when the main
826 daemon process exits with any of them, regardless of the
829 * systemctl's -H switch for connecting to remote systemd
830 machines has been extended so that it may be used to
831 directly connect to a specific container on the
832 host. "systemctl -H root@foobar:waldi" will now connect as
833 user "root" to host "foobar", and then proceed directly to
834 the container named "waldi". Note that currently you have to
835 authenticate as user "root" for this to work, as entering
836 containers is a privileged operation.
838 Contributions from: Andreas Henriksson, Benjamin Steinwender,
839 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
840 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
841 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
842 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
843 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
844 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
845 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
846 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
847 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
848 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
849 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
851 -- Berlin, 2014-07-03
855 * As an experimental feature, udev now tries to lock the
856 disk device node (flock(LOCK_SH|LOCK_NB)) while it
857 executes events for the disk or any of its partitions.
858 Applications like partitioning programs can lock the
859 disk device node (flock(LOCK_EX)) and claim temporary
860 device ownership that way; udev will entirely skip all event
861 handling for this disk and its partitions. If the disk
862 was opened for writing, the close will trigger a partition
863 table rescan in udev's "watch" facility, and if needed
864 synthesize "change" events for the disk and all its partitions.
865 This is now unconditionally enabled, and if it turns out to
866 cause major problems, we might turn it on only for specific
867 devices, or might need to disable it entirely. Device Mapper
868 devices are excluded from this logic.
870 * We temporarily dropped the "-l" switch for fsck invocations,
871 since they collide with the flock() logic above. util-linux
872 upstream has been changed already to avoid this conflict,
873 and we will readd "-l" as soon as util-linux with this
874 change has been released.
876 * The dependency on libattr has been removed. Since a long
877 time, the extended attribute calls have moved to glibc, and
878 libattr is thus unnecessary.
880 * Virtualization detection works without priviliges now. This
881 means the systemd-detect-virt binary no longer requires
882 CAP_SYS_PTRACE file capabilities, and our daemons can run
883 with fewer privileges.
885 * systemd-networkd now runs under its own "systemd-network"
886 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
887 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
888 loses the ability to write to files owned by root this way.
890 * Similar, systemd-resolved now runs under its own
891 "systemd-resolve" user with no capabilities remaining.
893 * Similar, systemd-bus-proxyd now runs under its own
894 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
896 * systemd-networkd gained support for setting up "veth"
897 virtual ethernet devices for container connectivity, as well
898 as GRE and VTI tunnels.
900 * systemd-networkd will no longer automatically attempt to
901 manually load kernel modules necessary for certain tunnel
902 transports. Instead, it is assumed the kernel loads them
903 automatically when required. This only works correctly on
904 very new kernels. On older kernels, please consider adding
905 the kernel modules to /etc/modules-load.d/ as a work-around.
907 * The resolv.conf file systemd-resolved generates has been
908 moved to /run/systemd/resolve/. If you have a symlink from
909 /etc/resolv.conf, it might be necessary to correct it.
911 * Two new service settings, ProtectHome= and ProtectSystem=,
912 have been added. When enabled, they will make the user data
913 (such as /home) inaccessible or read-only and the system
914 (such as /usr) read-only, for specific services. This allows
915 very light-weight per-service sandboxing to avoid
916 modifications of user data or system files from
917 services. These two new switches have been enabled for all
918 of systemd's long-running services, where appropriate.
920 * Socket units gained new SocketUser= and SocketGroup=
921 settings to set the owner user and group of AF_UNIX sockets
922 and FIFOs in the file system.
924 * Socket units gained a new RemoveOnStop= setting. If enabled,
925 all FIFOS and sockets in the file system will be removed
926 when the specific socket unit is stopped.
928 * Socket units gained a new Symlinks= setting. It takes a list
929 of symlinks to create to file system sockets or FIFOs
930 created by the specific Unix sockets. This is useful to
931 manage symlinks to socket nodes with the same life-cycle as
934 * The /dev/log socket and /dev/initctl FIFO have been moved to
935 /run, and have been replaced by symlinks. This allows
936 connecting to these facilities even if PrivateDevices=yes is
937 used for a service (which makes /dev/log itself unavailable,
938 but /run is left). This also has the benefit of ensuring
939 that /dev only contains device nodes, directories and
940 symlinks, and nothing else.
942 * sd-daemon gained two new calls sd_pid_notify() and
943 sd_pid_notifyf(). They are similar to sd_notify() and
944 sd_notifyf(), but allow overriding of the source PID of
945 notification messages if permissions permit this. This is
946 useful to send notify messages on behalf of a different
947 process (for example, the parent process). The
948 systemd-notify tool has been updated to make use of this
949 when sending messages (so that notification messages now
950 originate from the shell script invoking systemd-notify and
951 not the systemd-notify process itself. This should minimize
952 a race where systemd fails to associate notification
953 messages to services when the originating process already
956 * A new "on-abnormal" setting for Restart= has been added. If
957 set, it will result in automatic restarts on all "abnormal"
958 reasons for a process to exit, which includes unclean
959 signals, core dumps, timeouts and watchdog timeouts, but
960 does not include clean and unclean exit codes or clean
961 signals. Restart=on-abnormal is an alternative for
962 Restart=on-failure for services that shall be able to
963 terminate and avoid restarts on certain errors, by
964 indicating so with an unclean exit code. Restart=on-failure
965 or Restart=on-abnormal is now the recommended setting for
966 all long-running services.
968 * If the InaccessibleDirectories= service setting points to a
969 mount point (or if there are any submounts contained within
970 it), it is now attempted to completely unmount it, to make
971 the file systems truly unavailable for the respective
974 * The ReadOnlyDirectories= service setting and
975 systemd-nspawn's --read-only parameter are now recursively
976 applied to all submounts, too.
978 * Mount units may now be created transiently via the bus APIs.
980 * The support for SysV and LSB init scripts has been removed
981 from the systemd daemon itself. Instead, it is now
982 implemented as a generator that creates native systemd units
983 from these scripts when needed. This enables us to remove a
984 substantial amount of legacy code from PID 1, following the
985 fact that many distributions only ship a very small number
986 of LSB/SysV init scripts nowadays.
988 * Privileged Xen (dom0) domains are not considered
989 virtualization anymore by the virtualization detection
990 logic. After all, they generally have unrestricted access to
991 the hardware and usually are used to manage the unprivileged
994 * systemd-tmpfiles gained a new "C" line type, for copying
995 files or entire directories.
997 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
998 lines. So far, they have been non-globbing versions of the
999 latter, and have thus been redundant. In future, it is
1000 recommended to only use "z". "m" has hence been removed
1001 from the documentation, even though it stays supported.
1003 * A tmpfiles snippet to recreate the most basic structure in
1004 /var has been added. This is enough to create the /var/run →
1005 /run symlink and create a couple of structural
1006 directories. This allows systems to boot up with an empty or
1007 volatile /var. Of course, while with this change, the core OS
1008 now is capable with dealing with a volatile /var, not all
1009 user services are ready for it. However, we hope that sooner
1010 or later, many service daemons will be changed upstream so
1011 that they are able to automatically create their necessary
1012 directories in /var at boot, should they be missing. This is
1013 the first step to allow state-less systems that only require
1014 the vendor image for /usr to boot.
1016 * systemd-nspawn has gained a new --tmpfs= switch to mount an
1017 empty tmpfs instance to a specific directory. This is
1018 particularly useful for making use of the automatic
1019 reconstruction of /var (see above), by passing --tmpfs=/var.
1021 * Access modes specified in tmpfiles snippets may now be
1022 prefixed with "~", which indicates that they shall be masked
1023 by whether the existing file or directory is currently
1024 writable, readable or executable at all. Also, if specified,
1025 the sgid/suid/sticky bits will be masked for all
1028 * A new passive target unit "network-pre.target" has been
1029 added which is useful for services that shall run before any
1030 network is configured, for example firewall scripts.
1032 * The "floppy" group that previously owned the /dev/fd*
1033 devices is no longer used. The "disk" group is now used
1034 instead. Distributions should probably deprecate usage of
1037 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
1038 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
1039 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
1040 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
1041 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
1042 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
1045 -- Berlin, 2014-06-11
1049 * A new "systemd-timesyncd" daemon has been added for
1050 synchronizing the system clock across the network. It
1051 implements an SNTP client. In contrast to NTP
1052 implementations such as chrony or the NTP reference server,
1053 this only implements a client side, and does not bother with
1054 the full NTP complexity, focusing only on querying time from
1055 one remote server and synchronizing the local clock to
1056 it. Unless you intend to serve NTP to networked clients or
1057 want to connect to local hardware clocks, this simple NTP
1058 client should be more than appropriate for most
1059 installations. The daemon runs with minimal privileges, and
1060 has been hooked up with networkd to only operate when
1061 network connectivity is available. The daemon saves the
1062 current clock to disk every time a new NTP sync has been
1063 acquired, and uses this to possibly correct the system clock
1064 early at bootup, in order to accommodate for systems that
1065 lack an RTC such as the Raspberry Pi and embedded devices,
1066 and to make sure that time monotonically progresses on these
1067 systems, even if it is not always correct. To make use of
1068 this daemon, a new system user and group "systemd-timesync"
1069 needs to be created on installation of systemd.
1071 * The queue "seqnum" interface of libudev has been disabled, as
1072 it was generally incompatible with device namespacing as
1073 sequence numbers of devices go "missing" if the devices are
1074 part of a different namespace.
1076 * "systemctl list-timers" and "systemctl list-sockets" gained
1077 a --recursive switch for showing units of these types also
1078 for all local containers, similar in style to the already
1079 supported --recursive switch for "systemctl list-units".
1081 * A new RebootArgument= setting has been added for service
1082 units, which may be used to specify a kernel reboot argument
1083 to use when triggering reboots with StartLimitAction=.
1085 * A new FailureAction= setting has been added for service
1086 units which may be used to specify an operation to trigger
1087 when a service fails. This works similarly to
1088 StartLimitAction=, but unlike it, controls what is done
1089 immediately rather than only after several attempts to
1090 restart the service in question.
1092 * hostnamed got updated to also expose the kernel name,
1093 release, and version on the bus. This is useful for
1094 executing commands like hostnamectl with the -H switch.
1095 systemd-analyze makes use of this to properly display
1096 details when running non-locally.
1098 * The bootchart tool can now show cgroup information in the
1099 graphs it generates.
1101 * The CFS CPU quota cgroup attribute is now exposed for
1102 services. The new CPUQuota= switch has been added for this
1103 which takes a percentage value. Setting this will have the
1104 result that a service may never get more CPU time than the
1105 specified percentage, even if the machine is otherwise idle.
1107 * systemd-networkd learned IPIP and SIT tunnel support.
1109 * LSB init scripts exposing a dependency on $network will now
1110 get a dependency on network-online.target rather than simply
1111 network.target. This should bring LSB handling closer to
1112 what it was on SysV systems.
1114 * A new fsck.repair= kernel option has been added to control
1115 how fsck shall deal with unclean file systems at boot.
1117 * The (.ini) configuration file parser will now silently
1118 ignore sections whose name begins with "X-". This may be
1119 used to maintain application-specific extension sections in unit
1122 * machined gained a new API to query the IP addresses of
1123 registered containers. "machinectl status" has been updated
1124 to show these addresses in its output.
1126 * A new call sd_uid_get_display() has been added to the
1127 sd-login APIs for querying the "primary" session of a
1128 user. The "primary" session of the user is elected from the
1129 user's sessions and generally a graphical session is
1130 preferred over a text one.
1132 * A minimal systemd-resolved daemon has been added. It
1133 currently simply acts as a companion to systemd-networkd and
1134 manages resolv.conf based on per-interface DNS
1135 configuration, possibly supplied via DHCP. In the long run
1136 we hope to extend this into a local DNSSEC enabled DNS and
1139 * The systemd-networkd-wait-online tool is now enabled by
1140 default. It will delay network-online.target until a network
1141 connection has been configured. The tool primarily integrates
1142 with networkd, but will also make a best effort to make sense
1143 of network configuration performed in some other way.
1145 * Two new service options StartupCPUShares= and
1146 StartupBlockIOWeight= have been added that work similarly to
1147 CPUShares= and BlockIOWeight= however only apply during
1148 system startup. This is useful to prioritize certain services
1149 differently during bootup than during normal runtime.
1151 * hostnamed has been changed to prefer the statically
1152 configured hostname in /etc/hostname (unless set to
1153 'localhost' or empty) over any dynamic one supplied by
1154 dhcp. With this change, the rules for picking the hostname
1155 match more closely the rules of other configuration settings
1156 where the local administrator's configuration in /etc always
1157 overrides any other settings.
1159 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
1160 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
1161 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
1162 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
1163 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
1164 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
1165 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
1166 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
1167 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
1168 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
1169 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
1170 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
1171 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
1172 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
1173 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
1174 Lindskog, WaLyong Cho, Will Woods, Zbigniew
1177 -- Beijing, 2014-05-28
1181 * When restoring the screen brightness at boot, stay away from
1182 the darkest setting or from the lowest 5% of the available
1183 range, depending on which is the larger value of both. This
1184 should effectively protect the user from rebooting into a
1185 black screen, should the brightness have been set to minimum
1188 * sd-login gained a new sd_machine_get_class() call to
1189 determine the class ("vm" or "container") of a machine
1190 registered with machined.
1192 * sd-login gained new calls
1193 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
1194 to query the identity of the peer of a local AF_UNIX
1195 connection. They operate similarly to their sd_pid_get_xyz()
1198 * PID 1 will now maintain a system-wide system state engine
1199 with the states "starting", "running", "degraded",
1200 "maintenance", "stopping". These states are bound to system
1201 startup, normal runtime, runtime with at least one failed
1202 service, rescue/emergency mode and system shutdown. This
1203 state is shown in the "systemctl status" output when no unit
1204 name is passed. It is useful to determine system state, in
1205 particularly when doing so for many systems or containers at
1208 * A new command "list-machines" has been added to "systemctl"
1209 that lists all local OS containers and shows their system
1210 state (see above), if systemd runs inside of them.
1212 * systemctl gained a new "-r" switch to recursively enumerate
1213 units on all local containers, when used with the
1214 "list-unit" command (which is the default one that is
1215 executed when no parameters are specified).
1217 * The GPT automatic partition discovery logic will now honour
1218 two GPT partition flags: one may be set on a partition to
1219 cause it to be mounted read-only, and the other may be set
1220 on a partition to ignore it during automatic discovery.
1222 * Two new GPT type UUIDs have been added for automatic root
1223 partition discovery, for 32-bit and 64-bit ARM. This is not
1224 particularly useful for discovering the root directory on
1225 these architectures during bare-metal boots (since UEFI is
1226 not common there), but still very useful to allow booting of
1227 ARM disk images in nspawn with the -i option.
1229 * MAC addresses of interfaces created with nspawn's
1230 --network-interface= switch will now be generated from the
1231 machine name, and thus be stable between multiple invocations
1234 * logind will now automatically remove all IPC objects owned
1235 by a user if she or he fully logs out. This makes sure that
1236 users who are logged out cannot continue to consume IPC
1237 resources. This covers SysV memory, semaphores and message
1238 queues as well as POSIX shared memory and message
1239 queues. Traditionally, SysV and POSIX IPC had no life-cycle
1240 limits. With this functionality, that is corrected. This may
1241 be turned off by using the RemoveIPC= switch of logind.conf.
1243 * The systemd-machine-id-setup and tmpfiles tools gained a
1244 --root= switch to operate on a specific root directory,
1247 * journald can now forward logged messages to the TTYs of all
1248 logged in users ("wall"). This is the default for all
1249 emergency messages now.
1251 * A new tool systemd-journal-remote has been added to stream
1252 journal log messages across the network.
1254 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
1255 controller trees are mounted into it. Note that the
1256 directories mounted beneath it are not read-only. This is a
1257 security measure and is particularly useful because glibc
1258 actually includes a search logic to pick any tmpfs it can
1259 find to implement shm_open() if /dev/shm is not available
1260 (which it might very well be in namespaced setups).
1262 * machinectl gained a new "poweroff" command to cleanly power
1263 down a local OS container.
1265 * The PrivateDevices= unit file setting will now also drop the
1266 CAP_MKNOD capability from the capability bound set, and
1267 imply DevicePolicy=closed.
1269 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
1270 comprehensively on all long-running systemd services where
1271 this is appropriate.
1273 * systemd-udevd will now run in a disassociated mount
1274 namespace. To mount directories from udev rules, make sure to
1275 pull in mount units via SYSTEMD_WANTS properties.
1277 * The kdbus support gained support for uploading policy into
1278 the kernel. sd-bus gained support for creating "monitoring"
1279 connections that can eavesdrop into all bus communication
1280 for debugging purposes.
1282 * Timestamps may now be specified in seconds since the UNIX
1283 epoch Jan 1st, 1970 by specifying "@" followed by the value
1286 * Native tcpwrap support in systemd has been removed. tcpwrap
1287 is old code, not really maintained anymore and has serious
1288 shortcomings, and better options such as firewalls
1289 exist. For setups that require tcpwrap usage, please
1290 consider invoking your socket-activated service via tcpd,
1291 like on traditional inetd.
1293 * A new system.conf configuration option
1294 DefaultTimerAccuracySec= has been added that controls the
1295 default AccuracySec= setting of .timer units.
1297 * Timer units gained a new WakeSystem= switch. If enabled,
1298 timers configured this way will cause the system to resume
1299 from system suspend (if the system supports that, which most
1302 * Timer units gained a new Persistent= switch. If enabled,
1303 timers configured this way will save to disk when they have
1304 been last triggered. This information is then used on next
1305 reboot to possible execute overdue timer events, that
1306 could not take place because the system was powered off.
1307 This enables simple anacron-like behaviour for timer units.
1309 * systemctl's "list-timers" will now also list the time a
1310 timer unit was last triggered in addition to the next time
1311 it will be triggered.
1313 * systemd-networkd will now assign predictable IPv4LL
1314 addresses to its local interfaces.
1316 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
1317 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
1318 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
1319 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
1320 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
1321 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
1322 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
1323 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
1326 -- Berlin, 2014-03-25
1330 * A new unit file setting RestrictAddressFamilies= has been
1331 added to restrict which socket address families unit
1332 processes gain access to. This takes address family names
1333 like "AF_INET" or "AF_UNIX", and is useful to minimize the
1334 attack surface of services via exotic protocol stacks. This
1335 is built on seccomp system call filters.
1337 * Two new unit file settings RuntimeDirectory= and
1338 RuntimeDirectoryMode= have been added that may be used to
1339 manage a per-daemon runtime directories below /run. This is
1340 an alternative for setting up directory permissions with
1341 tmpfiles snippets, and has the advantage that the runtime
1342 directory's lifetime is bound to the daemon runtime and that
1343 the daemon starts up with an empty directory each time. This
1344 is particularly useful when writing services that drop
1345 privileges using the User= or Group= setting.
1347 * The DeviceAllow= unit setting now supports globbing for
1348 matching against device group names.
1350 * The systemd configuration file system.conf gained new
1351 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
1352 DefaultMemoryAccounting= to globally turn on/off accounting
1353 for specific resources (cgroups) for all units. These
1354 settings may still be overridden individually in each unit
1357 * systemd-gpt-auto-generator is now able to discover /srv and
1358 root partitions in addition to /home and swap partitions. It
1359 also supports LUKS-encrypted partitions now. With this in
1360 place, automatic discovery of partitions to mount following
1361 the Discoverable Partitions Specification
1362 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
1363 is now a lot more complete. This allows booting without
1364 /etc/fstab and without root= on the kernel command line on
1365 systems prepared appropriately.
1367 * systemd-nspawn gained a new --image= switch which allows
1368 booting up disk images and Linux installations on any block
1369 device that follow the Discoverable Partitions Specification
1370 (see above). This means that installations made with
1371 appropriately updated installers may now be started and
1372 deployed using container managers, completely
1373 unmodified. (We hope that libvirt-lxc will add support for
1374 this feature soon, too.)
1376 * systemd-nspawn gained a new --network-macvlan= setting to
1377 set up a private macvlan interface for the
1378 container. Similarly, systemd-networkd gained a new
1379 Kind=macvlan setting in .netdev files.
1381 * systemd-networkd now supports configuring local addresses
1384 * A new tool systemd-network-wait-online has been added to
1385 synchronously wait for network connectivity using
1388 * The sd-bus.h bus API gained a new sd_bus_track object for
1389 tracking the life-cycle of bus peers. Note that sd-bus.h is
1390 still not a public API though (unless you specify
1391 --enable-kdbus on the configure command line, which however
1392 voids your warranty and you get no API stability guarantee).
1394 * The $XDG_RUNTIME_DIR runtime directories for each user are
1395 now individual tmpfs instances, which has the benefit of
1396 introducing separate pools for each user, with individual
1397 size limits, and thus making sure that unprivileged clients
1398 can no longer negatively impact the system or other users by
1399 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
1400 RuntimeDirectorySize= has been introduced that allows
1401 controlling the default size limit for all users. It
1402 defaults to 10% of the available physical memory. This is no
1403 replacement for quotas on tmpfs though (which the kernel
1404 still does not support), as /dev/shm and /tmp are still
1405 shared resources used by both the system and unprivileged
1408 * logind will now automatically turn off automatic suspending
1409 on laptop lid close when more than one display is
1410 connected. This was previously expected to be implemented
1411 individually in desktop environments (such as GNOME),
1412 however has been added to logind now, in order to fix a
1413 boot-time race where a desktop environment might not have
1414 been started yet and thus not been able to take an inhibitor
1415 lock at the time where logind already suspends the system
1416 due to a closed lid.
1418 * logind will now wait at least 30s after each system
1419 suspend/resume cycle, and 3min after system boot before
1420 suspending the system due to a closed laptop lid. This
1421 should give USB docking stations and similar enough time to
1422 be probed and configured after system resume and boot in
1423 order to then act as suspend blocker.
1425 * systemd-run gained a new --property= setting which allows
1426 initialization of resource control properties (and others)
1427 for the created scope or service unit. Example: "systemd-run
1428 --property=BlockIOWeight=10 updatedb" may be used to run
1429 updatedb at a low block IO scheduling weight.
1431 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
1432 now also work in --scope mode.
1434 * When systemd is compiled with kdbus support, basic support
1435 for enforced policies is now in place. (Note that enabling
1436 kdbus still voids your warranty and no API compatibility
1439 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
1440 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1441 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
1442 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
1443 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
1444 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
1445 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
1446 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
1447 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
1448 Zbigniew Jędrzejewski-Szmek
1450 -- Berlin, 2014-03-12
1454 * systemd will now relabel /dev after loading the SMACK policy
1455 according to SMACK rules.
1457 * A new unit file option AppArmorProfile= has been added to
1458 set the AppArmor profile for the processes of a unit.
1460 * A new condition check ConditionArchitecture= has been added
1461 to conditionalize units based on the system architecture, as
1462 reported by uname()'s "machine" field.
1464 * systemd-networkd now supports matching on the system
1465 virtualization, architecture, kernel command line, host name
1468 * logind is now a lot more aggressive when suspending the
1469 machine due to a closed laptop lid. Instead of acting only
1470 on the lid close action, it will continuously watch the lid
1471 status and act on it. This is useful for laptops where the
1472 power button is on the outside of the chassis so that it can
1473 be reached without opening the lid (such as the Lenovo
1474 Yoga). On those machines, logind will now immediately
1475 re-suspend the machine if the power button has been
1476 accidentally pressed while the laptop was suspended and in a
1477 backpack or similar.
1479 * logind will now watch SW_DOCK switches and inhibit reaction
1480 to the lid switch if it is pressed. This means that logind
1481 will not suspend the machine anymore if the lid is closed
1482 and the system is docked, if the laptop supports SW_DOCK
1483 notifications via the input layer. Note that ACPI docking
1484 stations do not generate this currently. Also note that this
1485 logic is usually not fully sufficient and Desktop
1486 Environments should take a lid switch inhibitor lock when an
1487 external display is connected, as systemd will not watch
1490 * nspawn will now make use of the devices cgroup controller by
1491 default, and only permit creation of and access to the usual
1492 API device nodes like /dev/null or /dev/random, as well as
1493 access to (but not creation of) the pty devices.
1495 * We will now ship a default .network file for
1496 systemd-networkd that automatically configures DHCP for
1497 network interfaces created by nspawn's --network-veth or
1498 --network-bridge= switches.
1500 * systemd will now understand the usual M, K, G, T suffixes
1501 according to SI conventions (i.e. to the base 1000) when
1502 referring to throughput and hardware metrics. It will stay
1503 with IEC conventions (i.e. to the base 1024) for software
1504 metrics, according to what is customary according to
1505 Wikipedia. We explicitly document which base applies for
1506 each configuration option.
1508 * The DeviceAllow= setting in unit files now supports a syntax
1509 to whitelist an entire group of devices node majors at once,
1510 based on the /proc/devices listing. For example, with the
1511 string "char-pts", it is now possible to whitelist all
1512 current and future pseudo-TTYs at once.
1514 * sd-event learned a new "post" event source. Event sources of
1515 this type are triggered by the dispatching of any event
1516 source of a type that is not "post". This is useful for
1517 implementing clean-up and check event sources that are
1518 triggered by other work being done in the program.
1520 * systemd-networkd is no longer statically enabled, but uses
1521 the usual [Install] sections so that it can be
1522 enabled/disabled using systemctl. It still is enabled by
1525 * When creating a veth interface pair with systemd-nspawn, the
1526 host side will now be prefixed with "vb-" if
1527 --network-bridge= is used, and with "ve-" if --network-veth
1528 is used. This way, it is easy to distinguish these cases on
1529 the host, for example to apply different configuration to
1530 them with systemd-networkd.
1532 * The compatibility libraries for libsystemd-journal.so,
1533 libsystem-id128.so, libsystemd-login.so and
1534 libsystemd-daemon.so do not make use of IFUNC
1535 anymore. Instead, we now build libsystemd.so multiple times
1536 under these alternative names. This means that the footprint
1537 is drastically increased, but given that these are
1538 transitional compatibility libraries, this should not matter
1539 much. This change has been made necessary to support the ARM
1540 platform for these compatibility libraries, as the ARM
1541 toolchain is not really at the same level as the toolchain
1542 for other architectures like x86 and does not support
1543 IFUNC. Please make sure to use --enable-compat-libs only
1544 during a transitional period!
1546 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
1547 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1548 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
1549 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
1550 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
1551 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
1552 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
1553 Zbigniew Jędrzejewski-Szmek
1555 -- Berlin, 2014-02-24
1559 * A new component "systemd-networkd" has been added that can
1560 be used to configure local network interfaces statically or
1561 via DHCP. It is capable of bringing up bridges, VLANs, and
1562 bonding. Currently, no hook-ups for interactive network
1563 configuration are provided. Use this for your initrd,
1564 container, embedded, or server setup if you need a simple,
1565 yet powerful, network configuration solution. This
1566 configuration subsystem is quite nifty, as it allows wildcard
1567 hotplug matching in interfaces. For example, with a single
1568 configuration snippet, you can configure that all Ethernet
1569 interfaces showing up are automatically added to a bridge,
1570 or similar. It supports link-sensing and more.
1572 * A new tool "systemd-socket-proxyd" has been added which can
1573 act as a bidirectional proxy for TCP sockets. This is
1574 useful for adding socket activation support to services that
1575 do not actually support socket activation, including virtual
1576 machines and the like.
1578 * Add a new tool to save/restore rfkill state on
1581 * Save/restore state of keyboard backlights in addition to
1582 display backlights on shutdown/boot.
1584 * udev learned a new SECLABEL{} construct to label device
1585 nodes with a specific security label when they appear. For
1586 now, only SECLABEL{selinux} is supported, but the syntax is
1587 prepared for additional security frameworks.
1589 * udev gained a new scheme to configure link-level attributes
1590 from files in /etc/systemd/network/*.link. These files can
1591 match against MAC address, device path, driver name and type,
1592 and will apply attributes like the naming policy, link speed,
1593 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
1594 address assignment policy (randomized, ...).
1596 * The configuration of network interface naming rules for
1597 "permanent interface names" has changed: a new NamePolicy=
1598 setting in the [Link] section of .link files determines the
1599 priority of possible naming schemes (onboard, slot, mac,
1600 path). The default value of this setting is determined by
1601 /usr/lib/net/links/99-default.link. Old
1602 80-net-name-slot.rules udev configuration file has been
1603 removed, so local configuration overriding this file should
1604 be adapated to override 99-default.link instead.
1606 * When the User= switch is used in a unit file, also
1607 initialize $SHELL= based on the user database entry.
1609 * systemd no longer depends on libdbus. All communication is
1610 now done with sd-bus, systemd's low-level bus library
1613 * kdbus support has been added to PID 1 itself. When kdbus is
1614 enabled, this causes PID 1 to set up the system bus and
1615 enable support for a new ".busname" unit type that
1616 encapsulates bus name activation on kdbus. It works a little
1617 bit like ".socket" units, except for bus names. A new
1618 generator has been added that converts classic dbus1 service
1619 activation files automatically into native systemd .busname
1622 * sd-bus: add a light-weight vtable implementation that allows
1623 defining objects on the bus with a simple static const
1624 vtable array of its methods, signals and properties.
1626 * systemd will not generate or install static dbus
1627 introspection data anymore to /usr/share/dbus-1/interfaces,
1628 as the precise format of these files is unclear, and
1629 nothing makes use of it.
1631 * A proxy daemon is now provided to proxy clients connecting
1632 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
1633 compatibility with classic D-Bus.
1635 * A bus driver implementation has been added that supports the
1636 classic D-Bus bus driver calls on kdbus, also for
1637 compatibility purposes.
1639 * A new API "sd-event.h" has been added that implements a
1640 minimal event loop API built around epoll. It provides a
1641 couple of features that direct epoll usage is lacking:
1642 prioritization of events, scales to large numbers of timer
1643 events, per-event timer slack (accuracy), system-wide
1644 coalescing of timer events, exit handlers, watchdog
1645 supervision support using systemd's sd_notify() API, child
1648 * A new API "sd-rntl.h" has been added that provides an API
1649 around the route netlink interface of the kernel, similar in
1650 style to "sd-bus.h".
1652 * A new API "sd-dhcp-client.h" has been added that provides a
1653 small DHCPv4 client-side implementation. This is used by
1656 * There is a new kernel command line option
1657 "systemd.restore_state=0|1". When set to "0", none of the
1658 systemd tools will restore saved runtime state to hardware
1659 devices. More specifically, the rfkill and backlight states
1662 * The FsckPassNo= compatibility option in mount/service units
1663 has been removed. The fstab generator will now add the
1664 necessary dependencies automatically, and does not require
1665 PID1's support for that anymore.
1667 * journalctl gained a new switch, --list-boots, that lists
1668 recent boots with their times and boot IDs.
1670 * The various tools like systemctl, loginctl, timedatectl,
1671 busctl, systemd-run, ... have gained a new switch "-M" to
1672 connect to a specific, local OS container (as direct
1673 connection, without requiring SSH). This works on any
1674 container that is registered with machined, such as those
1675 created by libvirt-lxc or nspawn.
1677 * systemd-run and systemd-analyze also gained support for "-H"
1678 to connect to remote hosts via SSH. This is particularly
1679 useful for systemd-run because it enables queuing of jobs
1680 onto remote systems.
1682 * machinectl gained a new command "login" to open a getty
1683 login in any local container. This works with any container
1684 that is registered with machined (such as those created by
1685 libvirt-lxc or nspawn), and which runs systemd inside.
1687 * machinectl gained a new "reboot" command that may be used to
1688 trigger a reboot on a specific container that is registered
1689 with machined. This works on any container that runs an init
1690 system of some kind.
1692 * systemctl gained a new "list-timers" command to print a nice
1693 listing of installed timer units with the times they elapse
1696 * Alternative reboot() parameters may now be specified on the
1697 "systemctl reboot" command line and are passed to the
1698 reboot() system call.
1700 * systemctl gained a new --job-mode= switch to configure the
1701 mode to queue a job with. This is a more generic version of
1702 --fail, --irreversible, and --ignore-dependencies, which are
1703 still available but not advertised anymore.
1705 * /etc/systemd/system.conf gained new settings to configure
1706 various default timeouts of units, as well as the default
1707 start limit interval and burst. These may still be overridden
1710 * PID1 will now export on the bus profile data of the security
1711 policy upload process (such as the SELinux policy upload to
1714 * journald: when forwarding logs to the console, include
1715 timestamps (following the setting in
1716 /sys/module/printk/parameters/time).
1718 * OnCalendar= in timer units now understands the special
1719 strings "yearly" and "annually". (Both are equivalent)
1721 * The accuracy of timer units is now configurable with the new
1722 AccuracySec= setting. It defaults to 1min.
1724 * A new dependency type JoinsNamespaceOf= has been added that
1725 allows running two services within the same /tmp and network
1726 namespace, if PrivateNetwork= or PrivateTmp= are used.
1728 * A new command "cat" has been added to systemctl. It outputs
1729 the original unit file of a unit, and concatenates the
1730 contents of additional "drop-in" unit file snippets, so that
1731 the full configuration is shown.
1733 * systemctl now supports globbing on the various "list-xyz"
1734 commands, like "list-units" or "list-sockets", as well as on
1735 those commands which take multiple unit names.
1737 * journalctl's --unit= switch gained support for globbing.
1739 * All systemd daemons now make use of the watchdog logic so
1740 that systemd automatically notices when they hang.
1742 * If the $container_ttys environment variable is set,
1743 getty-generator will automatically spawn a getty for each
1744 listed tty. This is useful for container managers to request
1745 login gettys to be spawned on as many ttys as needed.
1747 * %h, %s, %U specifier support is not available anymore when
1748 used in unit files for PID 1. This is because NSS calls are
1749 not safe from PID 1. They stay available for --user
1750 instances of systemd, and as special case for the root user.
1752 * loginctl gained a new "--no-legend" switch to turn off output
1755 * The "sd-login.h" API gained three new calls:
1756 sd_session_is_remote(), sd_session_get_remote_user(),
1757 sd_session_get_remote_host() to query information about
1760 * The udev hardware database now also carries vendor/product
1761 information of SDIO devices.
1763 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
1764 determine whether watchdog notifications are requested by
1767 * Socket-activated per-connection services now include a
1768 short description of the connection parameters in the
1771 * tmpfiles gained a new "--boot" option. When this is not used,
1772 only lines where the command character is not suffixed with
1773 "!" are executed. When this option is specified, those
1774 options are executed too. This partitions tmpfiles
1775 directives into those that can be safely executed at any
1776 time, and those which should be run only at boot (for
1777 example, a line that creates /run/nologin).
1779 * A new API "sd-resolve.h" has been added which provides a simple
1780 asynchronous wrapper around glibc NSS host name resolution
1781 calls, such as getaddrinfo(). In contrast to glibc's
1782 getaddrinfo_a(), it does not use signals. In contrast to most
1783 other asynchronous name resolution libraries, this one does
1784 not reimplement DNS, but reuses NSS, so that alternate
1785 host name resolution systems continue to work, such as mDNS,
1786 LDAP, etc. This API is based on libasyncns, but it has been
1787 cleaned up for inclusion in systemd.
1789 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
1790 "sd-daemon.h" are no longer found in individual libraries
1791 libsystemd-journal.so, libsystemd-login.so,
1792 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
1793 merged them into a single library, libsystemd.so, which
1794 provides all symbols. The reason for this is cyclic
1795 dependencies, as these libraries tend to use each other's
1796 symbols. So far, we have managed to workaround that by linking
1797 a copy of a good part of our code into each of these
1798 libraries again and again, which, however, makes certain
1799 things hard to do, like sharing static variables. Also, it
1800 substantially increases footprint. With this change, there
1801 is only one library for the basic APIs systemd
1802 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
1803 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
1804 library as well, however are subject to the --enable-kdbus
1805 switch (see below). Note that "sd-dhcp-client.h" is not part
1806 of this library (this is because it only consumes, never
1807 provides, services of/to other APIs). To make the transition
1808 easy from the separate libraries to the unified one, we
1809 provide the --enable-compat-libs compile-time switch which
1810 will generate stub libraries that are compatible with the
1811 old ones but redirect all calls to the new one.
1813 * All of the kdbus logic and the new APIs "sd-bus.h",
1814 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
1815 and "sd-utf8.h" are compile-time optional via the
1816 "--enable-kdbus" switch, and they are not compiled in by
1817 default. To make use of kdbus, you have to explicitly enable
1818 the switch. Note however, that neither the kernel nor the
1819 userspace API for all of this is considered stable yet. We
1820 want to maintain the freedom to still change the APIs for
1821 now. By specifying this build-time switch, you acknowledge
1822 that you are aware of the instability of the current
1825 * Also, note that while kdbus is pretty much complete,
1826 it lacks one thing: proper policy support. This means you
1827 can build a fully working system with all features; however,
1828 it will be highly insecure. Policy support will be added in
1829 one of the next releases, at the same time that we will
1830 declare the APIs stable.
1832 * When the kernel command line argument "kdbus" is specified,
1833 systemd will automatically load the kdbus.ko kernel module. At
1834 this stage of development, it is only useful for testing kdbus
1835 and should not be used in production. Note: if "--enable-kdbus"
1836 is specified, and the kdbus.ko kernel module is available, and
1837 "kdbus" is added to the kernel command line, the entire system
1838 runs with kdbus instead of dbus-daemon, with the above mentioned
1839 problem of missing the system policy enforcement. Also a future
1840 version of kdbus.ko or a newer systemd will not be compatible with
1841 each other, and will unlikely be able to boot the machine if only
1842 one of them is updated.
1844 * systemctl gained a new "import-environment" command which
1845 uploads the caller's environment (or parts thereof) into the
1846 service manager so that it is inherited by services started
1847 by the manager. This is useful to upload variables like
1848 $DISPLAY into the user service manager.
1850 * A new PrivateDevices= switch has been added to service units
1851 which allows running a service with a namespaced /dev
1852 directory that does not contain any device nodes for
1853 physical devices. More specifically, it only includes devices
1854 such as /dev/null, /dev/urandom, and /dev/zero which are API
1857 * logind has been extended to support behaviour like VT
1858 switching on seats that do not support a VT. This makes
1859 multi-session available on seats that are not the first seat
1860 (seat0), and on systems where kernel support for VTs has
1861 been disabled at compile-time.
1863 * If a process holds a delay lock for system sleep or shutdown
1864 and fails to release it in time, we will now log its
1865 identity. This makes it easier to identify processes that
1866 cause slow suspends or power-offs.
1868 * When parsing /etc/crypttab, support for a new key-slot=
1869 option as supported by Debian is added. It allows indicating
1870 which LUKS slot to use on disk, speeding up key loading.
1872 * The sd_journald_sendv() API call has been checked and
1873 officially declared to be async-signal-safe so that it may
1874 be invoked from signal handlers for logging purposes.
1876 * Boot-time status output is now enabled automatically after a
1877 short timeout if boot does not progress, in order to give
1878 the user an indication what she or he is waiting for.
1880 * The boot-time output has been improved to show how much time
1881 remains until jobs expire.
1883 * The KillMode= switch in service units gained a new possible
1884 value "mixed". If set, and the unit is shut down, then the
1885 initial SIGTERM signal is sent only to the main daemon
1886 process, while the following SIGKILL signal is sent to
1887 all remaining processes of the service.
1889 * When a scope unit is registered, a new property "Controller"
1890 may be set. If set to a valid bus name, systemd will send a
1891 RequestStop() signal to this name when it would like to shut
1892 down the scope. This may be used to hook manager logic into
1893 the shutdown logic of scope units. Also, scope units may now
1894 be put in a special "abandoned" state, in which case the
1895 manager process which created them takes no further
1896 responsibilities for it.
1898 * When reading unit files, systemd will now verify
1899 the access mode of these files, and warn about certain
1900 suspicious combinations. This has been added to make it
1901 easier to track down packaging bugs where unit files are
1902 marked executable or world-writable.
1904 * systemd-nspawn gained a new "--setenv=" switch to set
1905 container-wide environment variables. The similar option in
1906 systemd-activate was renamed from "--environment=" to
1907 "--setenv=" for consistency.
1909 * systemd-nspawn has been updated to create a new kdbus domain
1910 for each container that is invoked, thus allowing each
1911 container to have its own set of system and user buses,
1912 independent of the host.
1914 * systemd-nspawn gained a new --drop-capability= switch to run
1915 the container with less capabilities than the default. Both
1916 --drop-capability= and --capability= now take the special
1917 string "all" for dropping or keeping all capabilities.
1919 * systemd-nspawn gained new switches for executing containers
1920 with specific SELinux labels set.
1922 * systemd-nspawn gained a new --quiet switch to not generate
1923 any additional output but the container's own console
1926 * systemd-nspawn gained a new --share-system switch to run a
1927 container without PID namespacing enabled.
1929 * systemd-nspawn gained a new --register= switch to control
1930 whether the container is registered with systemd-machined or
1931 not. This is useful for containers that do not run full
1932 OS images, but only specific apps.
1934 * systemd-nspawn gained a new --keep-unit which may be used
1935 when invoked as the only program from a service unit, and
1936 results in registration of the unit service itself in
1937 systemd-machined, instead of a newly opened scope unit.
1939 * systemd-nspawn gained a new --network-interface= switch for
1940 moving arbitrary interfaces to the container. The new
1941 --network-veth switch creates a virtual Ethernet connection
1942 between host and container. The new --network-bridge=
1943 switch then allows assigning the host side of this virtual
1944 Ethernet connection to a bridge device.
1946 * systemd-nspawn gained a new --personality= switch for
1947 setting the kernel personality for the container. This is
1948 useful when running a 32-bit container on a 64-bit host. A
1949 similar option Personality= is now also available for service
1952 * logind will now also track a "Desktop" identifier for each
1953 session which encodes the desktop environment of it. This is
1954 useful for desktop environments that want to identify
1955 multiple running sessions of itself easily.
1957 * A new SELinuxContext= setting for service units has been
1958 added that allows setting a specific SELinux execution
1959 context for a service.
1961 * Most systemd client tools will now honour $SYSTEMD_LESS for
1962 settings of the "less" pager. By default, these tools will
1963 override $LESS to allow certain operations to work, such as
1964 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
1965 influence this logic.
1967 * systemd's "seccomp" hook-up has been changed to make use of
1968 the libseccomp library instead of using its own
1969 implementation. This has benefits for portability among
1972 * For usage together with SystemCallFilter=, a new
1973 SystemCallErrorNumber= setting has been introduced that
1974 allows configuration of a system error number to be returned
1975 on filtered system calls, instead of immediately killing the
1976 process. Also, SystemCallArchitectures= has been added to
1977 limit access to system calls of a particular architecture
1978 (in order to turn off support for unused secondary
1979 architectures). There is also a global
1980 SystemCallArchitectures= setting in system.conf now to turn
1981 off support for non-native system calls system-wide.
1983 * systemd requires a kernel with a working name_to_handle_at(),
1984 please see the kernel config requirements in the README file.
1986 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
1987 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
1988 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
1989 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
1990 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
1991 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
1992 Elia Pinto, Florian Weimer, George McCollister, Goffredo
1993 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
1994 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
1995 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
1996 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
1997 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
1998 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
1999 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
2000 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
2001 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
2002 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
2003 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
2004 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
2005 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
2006 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
2007 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
2008 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
2009 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
2011 -- Berlin, 2014-02-20
2015 * logind has gained support for facilitating privileged input
2016 and drm device access for unprivileged clients. This work is
2017 useful to allow Wayland display servers (and similar
2018 programs, such as kmscon) to run under the user's ID and
2019 access input and drm devices which are normally
2020 protected. When this is used (and the kernel is new enough)
2021 logind will "mute" IO on the file descriptors passed to
2022 Wayland as long as it is in the background and "unmute" it
2023 if it returns into the foreground. This allows secure
2024 session switching without allowing background sessions to
2025 eavesdrop on input and display data. This also introduces
2026 session switching support if VT support is turned off in the
2027 kernel, and on seats that are not seat0.
2029 * A new kernel command line option luks.options= is understood
2030 now which allows specifying LUKS options for usage for LUKS
2031 encrypted partitions specified with luks.uuid=.
2033 * tmpfiles.d(5) snippets may now use specifier expansion in
2034 path names. More specifically %m, %b, %H, %v, are now
2035 replaced by the local machine id, boot id, hostname, and
2036 kernel version number.
2038 * A new tmpfiles.d(5) command "m" has been introduced which
2039 may be used to change the owner/group/access mode of a file
2040 or directory if it exists, but do nothing if it does not.
2042 * This release removes high-level support for the
2043 MemorySoftLimit= cgroup setting. The underlying kernel
2044 cgroup attribute memory.soft_limit= is currently badly
2045 designed and likely to be removed from the kernel API in its
2046 current form, hence we should not expose it for now.
2048 * The memory.use_hierarchy cgroup attribute is now enabled for
2049 all cgroups systemd creates in the memory cgroup
2050 hierarchy. This option is likely to be come the built-in
2051 default in the kernel anyway, and the non-hierarchical mode
2052 never made much sense in the intrinsically hierarchical
2055 * A new field _SYSTEMD_SLICE= is logged along with all journal
2056 messages containing the slice a message was generated
2057 from. This is useful to allow easy per-customer filtering of
2058 logs among other things.
2060 * systemd-journald will no longer adjust the group of journal
2061 files it creates to the "systemd-journal" group. Instead we
2062 rely on the journal directory to be owned by the
2063 "systemd-journal" group, and its setgid bit set, so that the
2064 kernel file system layer will automatically enforce that
2065 journal files inherit this group assignment. The reason for
2066 this change is that we cannot allow NSS look-ups from
2067 journald which would be necessary to resolve
2068 "systemd-journal" to a numeric GID, because this might
2069 create deadlocks if NSS involves synchronous queries to
2070 other daemons (such as nscd, or sssd) which in turn are
2071 logging clients of journald and might block on it, which
2072 would then dead lock. A tmpfiles.d(5) snippet included in
2073 systemd will make sure the setgid bit and group are
2074 properly set on the journal directory if it exists on every
2075 boot. However, we recommend adjusting it manually after
2076 upgrades too (or from RPM scriptlets), so that the change is
2077 not delayed until next reboot.
2079 * Backlight and random seed files in /var/lib/ have moved into
2080 the /var/lib/systemd/ directory, in order to centralize all
2081 systemd generated files in one directory.
2083 * Boot time performance measurements (as displayed by
2084 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
2085 performance information if that's available to determine how
2086 much time BIOS and boot loader initialization required. With
2087 a sufficiently new BIOS you hence no longer need to boot
2088 with Gummiboot to get access to such information.
2090 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
2091 Cristian Rodríguez, Dave Reisner, David Herrmann, David
2092 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
2093 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
2094 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
2095 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
2096 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2098 -- Berlin, 2013-10-02
2102 * The Restart= option for services now understands a new
2103 on-watchdog setting, which will restart the service
2104 automatically if the service stops sending out watchdog keep
2105 alive messages (as configured with WatchdogSec=).
2107 * The getty generator (which is responsible for bringing up a
2108 getty on configured serial consoles) will no longer only
2109 start a getty on the primary kernel console but on all
2110 others, too. This makes the order in which console= is
2111 specified on the kernel command line less important.
2113 * libsystemd-logind gained a new sd_session_get_vt() call to
2114 retrieve the VT number of a session.
2116 * If the option "tries=0" is set for an entry of /etc/crypttab
2117 its passphrase is queried indefinitely instead of any
2118 maximum number of tries.
2120 * If a service with a configure PID file terminates its PID
2121 file will now be removed automatically if it still exists
2122 afterwards. This should put an end to stale PID files.
2124 * systemd-run will now also take relative binary path names
2125 for execution and no longer insists on absolute paths.
2127 * InaccessibleDirectories= and ReadOnlyDirectories= now take
2128 paths that are optionally prefixed with "-" to indicate that
2129 it should not be considered a failure if they do not exist.
2131 * journalctl -o (and similar commands) now understands a new
2132 output mode "short-precise", it is similar to "short" but
2133 shows timestamps with usec accuracy.
2135 * The option "discard" (as known from Debian) is now
2136 synonymous to "allow-discards" in /etc/crypttab. In fact,
2137 "discard" is preferred now (since it is easier to remember
2140 * Some licensing clean-ups were made, so that more code is now
2141 LGPL-2.1 licensed than before.
2143 * A minimal tool to save/restore the display backlight
2144 brightness across reboots has been added. It will store the
2145 backlight setting as late as possible at shutdown, and
2146 restore it as early as possible during reboot.
2148 * A logic to automatically discover and enable home and swap
2149 partitions on GPT disks has been added. With this in place
2150 /etc/fstab becomes optional for many setups as systemd can
2151 discover certain partitions located on the root disk
2152 automatically. Home partitions are recognized under their
2153 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
2154 partitions are recognized under their GPT type ID
2155 0657fd6da4ab43c484e50933c84b4f4f.
2157 * systemd will no longer pass any environment from the kernel
2158 or initrd to system services. If you want to set an
2159 environment for all services, do so via the kernel command
2160 line systemd.setenv= assignment.
2162 * The systemd-sysctl tool no longer natively reads the file
2163 /etc/sysctl.conf. If desired, the file should be symlinked
2164 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
2165 legacy support by a symlink rather than built-in code, it
2166 also makes the otherwise hidden order of application of the
2167 different files visible. (Note that this partly reverts to a
2168 pre-198 application order of sysctl knobs!)
2170 * The "systemctl set-log-level" and "systemctl dump" commands
2171 have been moved to systemd-analyze.
2173 * systemd-run learned the new --remain-after-exit switch,
2174 which causes the scope unit not to be cleaned up
2175 automatically after the process terminated.
2177 * tmpfiles learned a new --exclude-prefix= switch to exclude
2178 certain paths from operation.
2180 * journald will now automatically flush all messages to disk
2181 as soon as a message at the log level CRIT, ALERT or EMERG
2184 Contributions from: Andrew Cook, Brandon Philips, Christian
2185 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
2186 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
2187 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
2188 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
2189 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
2190 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
2191 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
2192 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
2193 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
2194 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
2195 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
2196 William Giokas, Zbigniew Jędrzejewski-Szmek
2198 -- Berlin, 2013-09-13
2202 * The documentation has been updated to cover the various new
2203 concepts introduced with 205.
2205 * Unit files now understand the new %v specifier which
2206 resolves to the kernel version string as returned by "uname
2209 * systemctl now supports filtering the unit list output by
2210 load state, active state and sub state, using the new
2213 * "systemctl status" will now show the results of the
2214 condition checks (like ConditionPathExists= and similar) of
2215 the last start attempts of the unit. They are also logged to
2218 * "journalctl -b" may now be used to look for boot output of a
2219 specific boot. Try "journalctl -b -1" for the previous boot,
2220 but the syntax is substantially more powerful.
2222 * "journalctl --show-cursor" has been added which prints the
2223 cursor string the last shown log line. This may then be used
2224 with the new "journalctl --after-cursor=" switch to continue
2225 browsing logs from that point on.
2227 * "journalctl --force" may now be used to force regeneration
2230 * Creation of "dead" device nodes has been moved from udev
2231 into kmod and tmpfiles. Previously, udev would read the kmod
2232 databases to pre-generate dead device nodes based on meta
2233 information contained in kernel modules, so that these would
2234 be auto-loaded on access rather then at boot. As this
2235 does not really have much to do with the exposing actual
2236 kernel devices to userspace this has always been slightly
2237 alien in the udev codebase. Following the new scheme kmod
2238 will now generate a runtime snippet for tmpfiles from the
2239 module meta information and it now is tmpfiles' job to the
2240 create the nodes. This also allows overriding access and
2241 other parameters for the nodes using the usual tmpfiles
2242 facilities. As side effect this allows us to remove the
2243 CAP_SYS_MKNOD capability bit from udevd entirely.
2245 * logind's device ACLs may now be applied to these "dead"
2246 devices nodes too, thus finally allowing managed access to
2247 devices such as /dev/snd/sequencer whithout loading the
2248 backing module right-away.
2250 * A new RPM macro has been added that may be used to apply
2251 tmpfiles configuration during package installation.
2253 * systemd-detect-virt and ConditionVirtualization= now can
2254 detect User-Mode-Linux machines (UML).
2256 * journald will now implicitly log the effective capabilities
2257 set of processes in the message metadata.
2259 * systemd-cryptsetup has gained support for TrueCrypt volumes.
2261 * The initrd interface has been simplified (more specifically,
2262 support for passing performance data via environment
2263 variables and fsck results via files in /run has been
2264 removed). These features were non-essential, and are
2265 nowadays available in a much nicer way by having systemd in
2266 the initrd serialize its state and have the hosts systemd
2267 deserialize it again.
2269 * The udev "keymap" data files and tools to apply keyboard
2270 specific mappings of scan to key codes, and force-release
2271 scan code lists have been entirely replaced by a udev
2272 "keyboard" builtin and a hwdb data file.
2274 * systemd will now honour the kernel's "quiet" command line
2275 argument also during late shutdown, resulting in a
2276 completely silent shutdown when used.
2278 * There's now an option to control the SO_REUSEPORT socket
2279 option in .socket units.
2281 * Instance units will now automatically get a per-template
2282 subslice of system.slice unless something else is explicitly
2283 configured. For example, instances of sshd@.service will now
2284 implicitly be placed in system-sshd.slice rather than
2285 system.slice as before.
2287 * Test coverage support may now be enabled at build time.
2289 Contributions from: Dave Reisner, Frederic Crozat, Harald
2290 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
2291 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
2292 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
2293 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
2294 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
2295 Giokas, Zbigniew Jędrzejewski-Szmek
2297 -- Berlin, 2013-07-23
2301 * Two new unit types have been introduced:
2303 Scope units are very similar to service units, however, are
2304 created out of pre-existing processes -- instead of PID 1
2305 forking off the processes. By using scope units it is
2306 possible for system services and applications to group their
2307 own child processes (worker processes) in a powerful way
2308 which then maybe used to organize them, or kill them
2309 together, or apply resource limits on them.
2311 Slice units may be used to partition system resources in an
2312 hierarchical fashion and then assign other units to them. By
2313 default there are now three slices: system.slice (for all
2314 system services), user.slice (for all user sessions),
2315 machine.slice (for VMs and containers).
2317 Slices and scopes have been introduced primarily in
2318 context of the work to move cgroup handling to a
2319 single-writer scheme, where only PID 1
2320 creates/removes/manages cgroups.
2322 * There's a new concept of "transient" units. In contrast to
2323 normal units these units are created via an API at runtime,
2324 not from configuration from disk. More specifically this
2325 means it is now possible to run arbitrary programs as
2326 independent services, with all execution parameters passed
2327 in via bus APIs rather than read from disk. Transient units
2328 make systemd substantially more dynamic then it ever was,
2329 and useful as a general batch manager.
2331 * logind has been updated to make use of scope and slice units
2332 for managing user sessions. As a user logs in he will get
2333 his own private slice unit, to which all sessions are added
2334 as scope units. We also added support for automatically
2335 adding an instance of user@.service for the user into the
2336 slice. Effectively logind will no longer create cgroup
2337 hierarchies on its own now, it will defer entirely to PID 1
2338 for this by means of scope, service and slice units. Since
2339 user sessions this way become entities managed by PID 1
2340 the output of "systemctl" is now a lot more comprehensive.
2342 * A new mini-daemon "systemd-machined" has been added which
2343 may be used by virtualization managers to register local
2344 VMs/containers. nspawn has been updated accordingly, and
2345 libvirt will be updated shortly. machined will collect a bit
2346 of meta information about the VMs/containers, and assign
2347 them their own scope unit (see above). The collected
2348 meta-data is then made available via the "machinectl" tool,
2349 and exposed in "ps" and similar tools. machined/machinectl
2350 is compile-time optional.
2352 * As discussed earlier, the low-level cgroup configuration
2353 options ControlGroup=, ControlGroupModify=,
2354 ControlGroupPersistent=, ControlGroupAttribute= have been
2355 removed. Please use high-level attribute settings instead as
2356 well as slice units.
2358 * A new bus call SetUnitProperties() has been added to alter
2359 various runtime parameters of a unit. This is primarily
2360 useful to alter cgroup parameters dynamically in a nice way,
2361 but will be extended later on to make more properties
2362 modifiable at runtime. systemctl gained a new set-properties
2363 command that wraps this call.
2365 * A new tool "systemd-run" has been added which can be used to
2366 run arbitrary command lines as transient services or scopes,
2367 while configuring a number of settings via the command
2368 line. This tool is currently very basic, however already
2369 very useful. We plan to extend this tool to even allow
2370 queuing of execution jobs with time triggers from the
2371 command line, similar in fashion to "at".
2373 * nspawn will now inform the user explicitly that kernels with
2374 audit enabled break containers, and suggest the user to turn
2377 * Support for detecting the IMA and AppArmor security
2378 frameworks with ConditionSecurity= has been added.
2380 * journalctl gained a new "-k" switch for showing only kernel
2381 messages, mimicking dmesg output; in addition to "--user"
2382 and "--system" switches for showing only user's own logs
2385 * systemd-delta can now show information about drop-in
2386 snippets extending unit files.
2388 * libsystemd-bus has been substantially updated but is still
2389 not available as public API.
2391 * systemd will now look for the "debug" argument on the kernel
2392 command line and enable debug logging, similar to what
2393 "systemd.log_level=debug" already did before.
2395 * "systemctl set-default", "systemctl get-default" has been
2396 added to configure the default.target symlink, which
2397 controls what to boot into by default.
2399 * "systemctl set-log-level" has been added as a convenient
2400 way to raise and lower systemd logging threshold.
2402 * "systemd-analyze plot" will now show the time the various
2403 generators needed for execution, as well as information
2404 about the unit file loading.
2406 * libsystemd-journal gained a new sd_journal_open_files() call
2407 for opening specific journal files. journactl also gained a
2408 new switch to expose this new functionality. Previously we
2409 only supported opening all files from a directory, or all
2410 files from the system, as opening individual files only is
2411 racy due to journal file rotation.
2413 * systemd gained the new DefaultEnvironment= setting in
2414 /etc/systemd/system.conf to set environment variables for
2417 * If a privileged process logs a journal message with the
2418 OBJECT_PID= field set, then journald will automatically
2419 augment this with additional OBJECT_UID=, OBJECT_GID=,
2420 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
2421 system services want to log events about specific client
2422 processes. journactl/systemctl has been updated to make use
2423 of this information if all log messages regarding a specific
2426 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
2427 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
2428 Reisner, David Coppa, David King, David Strauss, Eelco
2429 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
2430 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
2431 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
2432 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
2433 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
2434 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
2435 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
2436 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
2437 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
2438 Łukasz Stelmach, 장동준
2442 * The Python bindings gained some minimal support for the APIs
2443 exposed by libsystemd-logind.
2445 * ConditionSecurity= gained support for detecting SMACK. Since
2446 this condition already supports SELinux and AppArmor we only
2447 miss IMA for this. Patches welcome!
2449 Contributions from: Karol Lewandowski, Lennart Poettering,
2450 Zbigniew Jędrzejewski-Szmek
2454 * systemd-nspawn will now create /etc/resolv.conf if
2455 necessary, before bind-mounting the host's file onto it.
2457 * systemd-nspawn will now store meta information about a
2458 container on the container's cgroup as extended attribute
2459 fields, including the root directory.
2461 * The cgroup hierarchy has been reworked in many ways. All
2462 objects any of the components systemd creates in the cgroup
2463 tree are now suffixed. More specifically, user sessions are
2464 now placed in cgroups suffixed with ".session", users in
2465 cgroups suffixed with ".user", and nspawn containers in
2466 cgroups suffixed with ".nspawn". Furthermore, all cgroup
2467 names are now escaped in a simple scheme to avoid collision
2468 of userspace object names with kernel filenames. This work
2469 is preparation for making these objects relocatable in the
2470 cgroup tree, in order to allow easy resource partitioning of
2471 these objects without causing naming conflicts.
2473 * systemctl list-dependencies gained the new switches
2474 --plain, --reverse, --after and --before.
2476 * systemd-inhibit now shows the process name of processes that
2477 have taken an inhibitor lock.
2479 * nss-myhostname will now also resolve "localhost"
2480 implicitly. This makes /etc/hosts an optional file and
2481 nicely handles that on IPv6 ::1 maps to both "localhost" and
2484 * libsystemd-logind.so gained a new call
2485 sd_get_machine_names() to enumerate running containers and
2486 VMs (currently only supported by very new libvirt and
2487 nspawn). sd_login_monitor can now be used to watch
2488 VMs/containers coming and going.
2490 * .include is not allowed recursively anymore, and only in
2491 unit files. Usually it is better to use drop-in snippets in
2492 .d/*.conf anyway, as introduced with systemd 198.
2494 * systemd-analyze gained a new "critical-chain" command that
2495 determines the slowest chain of units run during system
2496 boot-up. It is very useful for tracking down where
2497 optimizing boot time is the most beneficial.
2499 * systemd will no longer allow manipulating service paths in
2500 the name=systemd:/system cgroup tree using ControlGroup= in
2501 units. (But is still fine with it in all other dirs.)
2503 * There's a new systemd-nspawn@.service service file that may
2504 be used to easily run nspawn containers as system
2505 services. With the container's root directory in
2506 /var/lib/container/foobar it is now sufficient to run
2507 "systemctl start systemd-nspawn@foobar.service" to boot it.
2509 * systemd-cgls gained a new parameter "--machine" to list only
2510 the processes within a certain container.
2512 * ConditionSecurity= now can check for "apparmor". We still
2513 are lacking checks for SMACK and IMA for this condition
2514 check though. Patches welcome!
2516 * A new configuration file /etc/systemd/sleep.conf has been
2517 added that may be used to configure which kernel operation
2518 systemd is supposed to execute when "suspend", "hibernate"
2519 or "hybrid-sleep" is requested. This makes the new kernel
2520 "freeze" state accessible to the user.
2522 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
2523 the passed argument if applicable.
2525 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
2526 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
2527 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
2528 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
2529 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
2530 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
2531 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
2536 * The output of 'systemctl list-jobs' got some polishing. The
2537 '--type=' argument may now be passed more than once. A new
2538 command 'systemctl list-sockets' has been added which shows
2539 a list of kernel sockets systemd is listening on with the
2540 socket units they belong to, plus the units these socket
2543 * The experimental libsystemd-bus library got substantial
2544 updates to work in conjunction with the (also experimental)
2545 kdbus kernel project. It works well enough to exchange
2546 messages with some sophistication. Note that kdbus is not
2547 ready yet, and the library is mostly an elaborate test case
2548 for now, and not installable.
2550 * systemd gained a new unit 'systemd-static-nodes.service'
2551 that generates static device nodes earlier during boot, and
2552 can run in conjunction with udev.
2554 * libsystemd-login gained a new call sd_pid_get_user_unit()
2555 to retrieve the user systemd unit a process is running
2556 in. This is useful for systems where systemd is used as
2559 * systemd-nspawn now places all containers in the new /machine
2560 top-level cgroup directory in the name=systemd
2561 hierarchy. libvirt will soon do the same, so that we get a
2562 uniform separation of /system, /user and /machine for system
2563 services, user processes and containers/virtual
2564 machines. This new cgroup hierarchy is also useful to stick
2565 stable names to specific container instances, which can be
2566 recognized later this way (this name may be controlled
2567 via systemd-nspawn's new -M switch). libsystemd-login also
2568 gained a new call sd_pid_get_machine_name() to retrieve the
2569 name of the container/VM a specific process belongs to.
2571 * bootchart can now store its data in the journal.
2573 * libsystemd-journal gained a new call
2574 sd_journal_add_conjunction() for AND expressions to the
2575 matching logic. This can be used to express more complex
2576 logical expressions.
2578 * journactl can now take multiple --unit= and --user-unit=
2581 * The cryptsetup logic now understands the "luks.key=" kernel
2582 command line switch for specifying a file to read the
2583 decryption key from. Also, if a configured key file is not
2584 found the tool will now automatically fall back to prompting
2587 * Python systemd.journal module was updated to wrap recently
2588 added functions from libsystemd-journal. The interface was
2589 changed to bring the low level interface in s.j._Reader
2590 closer to the C API, and the high level interface in
2591 s.j.Reader was updated to wrap and convert all data about
2594 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
2595 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
2596 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
2597 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
2598 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
2599 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2603 * journalctl --update-catalog now understands a new --root=
2604 option to operate on catalogs found in a different root
2607 * During shutdown after systemd has terminated all running
2608 services a final killing loop kills all remaining left-over
2609 processes. We will now print the name of these processes
2610 when we send SIGKILL to them, since this usually indicates a
2613 * If /etc/crypttab refers to password files stored on
2614 configured mount points automatic dependencies will now be
2615 generated to ensure the specific mount is established first
2616 before the key file is attempted to be read.
2618 * 'systemctl status' will now show information about the
2619 network sockets a socket unit is listening on.
2621 * 'systemctl status' will also shown information about any
2622 drop-in configuration file for units. (Drop-In configuration
2623 files in this context are files such as
2624 /etc/systemd/systemd/foobar.service.d/*.conf)
2626 * systemd-cgtop now optionally shows summed up CPU times of
2627 cgroups. Press '%' while running cgtop to switch between
2628 percentage and absolute mode. This is useful to determine
2629 which cgroups use up the most CPU time over the entire
2630 runtime of the system. systemd-cgtop has also been updated
2631 to be 'pipeable' for processing with further shell tools.
2633 * 'hostnamectl set-hostname' will now allow setting of FQDN
2636 * The formatting and parsing of time span values has been
2637 changed. The parser now understands fractional expressions
2638 such as "5.5h". The formatter will now output fractional
2639 expressions for all time spans under 1min, i.e. "5.123456s"
2640 rather than "5s 123ms 456us". For time spans under 1s
2641 millisecond values are shown, for those under 1ms
2642 microsecond values are shown. This should greatly improve
2643 all time-related output of systemd.
2645 * libsystemd-login and libsystemd-journal gained new
2646 functions for querying the poll() events mask and poll()
2647 timeout value for integration into arbitrary event
2650 * localectl gained the ability to list available X11 keymaps
2651 (models, layouts, variants, options).
2653 * 'systemd-analyze dot' gained the ability to filter for
2654 specific units via shell-style globs, to create smaller,
2655 more useful graphs. I.e. it is now possible to create simple
2656 graphs of all the dependencies between only target units, or
2657 of all units that Avahi has dependencies with.
2659 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
2660 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
2661 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
2662 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
2663 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
2664 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
2665 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
2669 * The boot-time readahead implementation for rotating media
2670 will now read the read-ahead data in multiple passes which
2671 consist of all read requests made in equidistant time
2672 intervals. This means instead of strictly reading read-ahead
2673 data in its physical order on disk we now try to find a
2674 middle ground between physical and access time order.
2676 * /etc/os-release files gained a new BUILD_ID= field for usage
2677 on operating systems that provide continuous builds of OS
2680 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
2681 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
2682 William Douglas, Zbigniew Jędrzejewski-Szmek
2686 * systemd-python gained an API exposing libsystemd-daemon.
2688 * The SMACK setup logic gained support for uploading CIPSO
2691 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
2692 ReadOnlyDirectories= and InaccessibleDirectories= has
2693 changed. The private /tmp and /var/tmp directories are now
2694 shared by all processes of a service (which means
2695 ExecStartPre= may now leave data in /tmp that ExecStart= of
2696 the same service can still access). When a service is
2697 stopped its temporary directories are immediately deleted
2698 (normal clean-up with tmpfiles is still done in addition to
2701 * By default, systemd will now set a couple of sysctl
2702 variables in the kernel: the safe sysrq options are turned
2703 on, IP route verification is turned on, and source routing
2704 disabled. The recently added hardlink and softlink
2705 protection of the kernel is turned on. These settings should
2706 be reasonably safe, and good defaults for all new systems.
2708 * The predictable network naming logic may now be turned off
2709 with a new kernel command line switch: net.ifnames=0.
2711 * A new libsystemd-bus module has been added that implements a
2712 pretty complete D-Bus client library. For details see:
2714 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
2716 * journald will now explicitly flush the journal files to disk
2717 at the latest 5min after each write. The file will then also
2718 be marked offline until the next write. This should increase
2719 reliability in case of a crash. The synchronization delay
2720 can be configured via SyncIntervalSec= in journald.conf.
2722 * There's a new remote-fs-setup.target unit that can be used
2723 to pull in specific services when at least one remote file
2724 system is to be mounted.
2726 * There are new targets timers.target and paths.target as
2727 canonical targets to pull user timer and path units in
2728 from. This complements sockets.target with a similar
2729 purpose for socket units.
2731 * libudev gained a new call udev_device_set_attribute_value()
2732 to set sysfs attributes of a device.
2734 * The udev daemon now sets the default number of worker
2735 processes executed in parallel based on the number of available
2736 CPUs instead of the amount of available RAM. This is supposed
2737 to provide a more reliable default and limit a too aggressive
2738 paralellism for setups with 1000s of devices connected.
2740 Contributions from: Auke Kok, Colin Walters, Cristian
2741 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
2742 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
2743 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
2744 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
2745 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
2746 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
2747 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
2748 Zbigniew Jędrzejewski-Szmek
2752 * Configuration of unit files may now be extended via drop-in
2753 files without having to edit/override the unit files
2754 themselves. More specifically, if the administrator wants to
2755 change one value for a service file foobar.service he can
2756 now do so by dropping in a configuration snippet into
2757 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
2758 will load all these snippets and apply them on top of the
2759 main unit configuration file, possibly extending or
2760 overriding its settings. Using these drop-in snippets is
2761 generally nicer than the two earlier options for changing
2762 unit files locally: copying the files from
2763 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
2764 them there; or creating a new file in /etc/systemd/system/
2765 that incorporates the original one via ".include". Drop-in
2766 snippets into these .d/ directories can be placed in any
2767 directory systemd looks for units in, and the usual
2768 overriding semantics between /usr/lib, /etc and /run apply
2771 * Most unit file settings which take lists of items can now be
2772 reset by assigning the empty string to them. For example,
2773 normally, settings such as Environment=FOO=BAR append a new
2774 environment variable assignment to the environment block,
2775 each time they are used. By assigning Environment= the empty
2776 string the environment block can be reset to empty. This is
2777 particularly useful with the .d/*.conf drop-in snippets
2778 mentioned above, since this adds the ability to reset list
2779 settings from vendor unit files via these drop-ins.
2781 * systemctl gained a new "list-dependencies" command for
2782 listing the dependencies of a unit recursively.
2784 * Inhibitors are now honored and listed by "systemctl
2785 suspend", "systemctl poweroff" (and similar) too, not only
2786 GNOME. These commands will also list active sessions by
2789 * Resource limits (as exposed by the various control group
2790 controllers) can now be controlled dynamically at runtime
2791 for all units. More specifically, you can now use a command
2792 like "systemctl set-cgroup-attr foobar.service cpu.shares
2793 2000" to alter the CPU shares a specific service gets. These
2794 settings are stored persistently on disk, and thus allow the
2795 administrator to easily adjust the resource usage of
2796 services with a few simple commands. This dynamic resource
2797 management logic is also available to other programs via the
2798 bus. Almost any kernel cgroup attribute and controller is
2801 * systemd-vconsole-setup will now copy all font settings to
2802 all allocated VTs, where it previously applied them only to
2805 * libsystemd-login gained the new sd_session_get_tty() API
2808 * This release drops support for a few legacy or
2809 distribution-specific LSB facility names when parsing init
2810 scripts: $x-display-manager, $mail-transfer-agent,
2811 $mail-transport-agent, $mail-transfer-agent, $smtp,
2812 $null. Also, the mail-transfer-agent.target unit backing
2813 this has been removed. Distributions which want to retain
2814 compatibility with this should carry the burden for
2815 supporting this themselves and patch support for these back
2816 in, if they really need to. Also, the facilities $syslog and
2817 $local_fs are now ignored, since systemd does not support
2818 early-boot LSB init scripts anymore, and these facilities
2819 are implied anyway for normal services. syslog.target has
2822 * There are new bus calls on PID1's Manager object for
2823 cancelling jobs, and removing snapshot units. Previously,
2824 both calls were only available on the Job and Snapshot
2827 * systemd-journal-gatewayd gained SSL support.
2829 * The various "environment" files, such as /etc/locale.conf
2830 now support continuation lines with a backslash ("\") as
2831 last character in the line, similarly in style (but different)
2832 to how this is supported in shells.
2834 * For normal user processes the _SYSTEMD_USER_UNIT= field is
2835 now implicitly appended to every log entry logged. systemctl
2836 has been updated to filter by this field when operating on a
2837 user systemd instance.
2839 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
2840 CAP_AUDIT_CONTROL capabilities to the capabilities set for
2841 the container. This makes it easier to boot unmodified
2842 Fedora systems in a container, which however still requires
2843 audit=0 to be passed on the kernel command line. Auditing in
2844 kernel and userspace is unfortunately still too broken in
2845 context of containers, hence we recommend compiling it out
2846 of the kernel or using audit=0. Hopefully this will be fixed
2847 one day for good in the kernel.
2849 * nspawn gained the new --bind= and --bind-ro= parameters to
2850 bind mount specific directories from the host into the
2853 * nspawn will now mount its own devpts file system instance
2854 into the container, in order not to leak pty devices from
2855 the host into the container.
2857 * systemd will now read the firmware boot time performance
2858 information from the EFI variables, if the used boot loader
2859 supports this, and takes it into account for boot performance
2860 analysis via "systemd-analyze". This is currently supported
2861 only in conjunction with Gummiboot, but could be supported
2862 by other boot loaders too. For details see:
2864 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
2866 * A new generator has been added that automatically mounts the
2867 EFI System Partition (ESP) to /boot, if that directory
2868 exists, is empty, and no other file system has been
2869 configured to be mounted there.
2871 * logind will now send out PrepareForSleep(false) out
2872 unconditionally, after coming back from suspend. This may be
2873 used by applications as asynchronous notification for
2874 system resume events.
2876 * "systemctl unlock-sessions" has been added, that allows
2877 unlocking the screens of all user sessions at once, similar
2878 to how "systemctl lock-sessions" already locked all users
2879 sessions. This is backed by a new D-Bus call UnlockSessions().
2881 * "loginctl seat-status" will now show the master device of a
2882 seat. (i.e. the device of a seat that needs to be around for
2883 the seat to be considered available, usually the graphics
2886 * tmpfiles gained a new "X" line type, that allows
2887 configuration of files and directories (with wildcards) that
2888 shall be excluded from automatic cleanup ("aging").
2890 * udev default rules set the device node permissions now only
2891 at "add" events, and do not change them any longer with a
2892 later "change" event.
2894 * The log messages for lid events and power/sleep keypresses
2895 now carry a message ID.
2897 * We now have a substantially larger unit test suite, but this
2898 continues to be work in progress.
2900 * udevadm hwdb gained a new --root= parameter to change the
2901 root directory to operate relative to.
2903 * logind will now issue a background sync() request to the kernel
2904 early at shutdown, so that dirty buffers are flushed to disk early
2905 instead of at the last moment, in order to optimize shutdown
2908 * A new bootctl tool has been added that is an interface for
2909 certain boot loader operations. This is currently a preview
2910 and is likely to be extended into a small mechanism daemon
2911 like timedated, localed, hostnamed, and can be used by
2912 graphical UIs to enumerate available boot options, and
2913 request boot into firmware operations.
2915 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
2916 the rest of the package. It also has been updated to work
2917 correctly in initrds.
2919 * Policykit previously has been runtime optional, and is now
2920 also compile time optional via a configure switch.
2922 * systemd-analyze has been reimplemented in C. Also "systemctl
2923 dot" has moved into systemd-analyze.
2925 * "systemctl status" with no further parameters will now print
2926 the status of all active or failed units.
2928 * Operations such as "systemctl start" can now be executed
2929 with a new mode "--irreversible" which may be used to queue
2930 operations that cannot accidentally be reversed by a later
2931 job queuing. This is by default used to make shutdown
2932 requests more robust.
2934 * The Python API of systemd now gained a new module for
2935 reading journal files.
2937 * A new tool kernel-install has been added that can install
2938 kernel images according to the Boot Loader Specification:
2940 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
2942 * Boot time console output has been improved to provide
2943 animated boot time output for hanging jobs.
2945 * A new tool systemd-activate has been added which can be used
2946 to test socket activation with, directly from the command
2947 line. This should make it much easier to test and debug
2948 socket activation in daemons.
2950 * journalctl gained a new "--reverse" (or -r) option to show
2951 journal output in reverse order (i.e. newest line first).
2953 * journalctl gained a new "--pager-end" (or -e) option to jump
2954 to immediately jump to the end of the journal in the
2955 pager. This is only supported in conjunction with "less".
2957 * journalctl gained a new "--user-unit=" option, that works
2958 similarly to "--unit=" but filters for user units rather than
2961 * A number of unit files to ease adoption of systemd in
2962 initrds has been added. This moves some minimal logic from
2963 the various initrd implementations into systemd proper.
2965 * The journal files are now owned by a new group
2966 "systemd-journal", which exists specifically to allow access
2967 to the journal, and nothing else. Previously, we used the
2968 "adm" group for that, which however possibly covers more
2969 than just journal/log file access. This new group is now
2970 already used by systemd-journal-gatewayd to ensure this
2971 daemon gets access to the journal files and as little else
2972 as possible. Note that "make install" will also set FS ACLs
2973 up for /var/log/journal to give "adm" and "wheel" read
2974 access to it, in addition to "systemd-journal" which owns
2975 the journal files. We recommend that packaging scripts also
2976 add read access to "adm" + "wheel" to /var/log/journal, and
2977 all existing/future journal files. To normal users and
2978 administrators little changes, however packagers need to
2979 ensure to create the "systemd-journal" system group at
2980 package installation time.
2982 * The systemd-journal-gatewayd now runs as unprivileged user
2983 systemd-journal-gateway:systemd-journal-gateway. Packaging
2984 scripts need to create these system user/group at
2987 * timedated now exposes a new boolean property CanNTP that
2988 indicates whether a local NTP service is available or not.
2990 * systemd-detect-virt will now also detect xen PVs
2992 * The pstore file system is now mounted by default, if it is
2995 * In addition to the SELinux and IMA policies we will now also
2996 load SMACK policies at early boot.
2998 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
2999 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
3000 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
3001 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
3002 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
3003 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
3004 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
3005 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
3006 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
3007 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
3008 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
3009 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
3010 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
3011 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
3015 * Timer units now support calendar time events in addition to
3016 monotonic time events. That means you can now trigger a unit
3017 based on a calendar time specification such as "Thu,Fri
3018 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
3019 or fifth day of any month of the year 2013, given that it is
3020 a thursday or friday. This brings timer event support
3021 considerably closer to cron's capabilities. For details on
3022 the supported calendar time specification language see
3025 * udev now supports a number of different naming policies for
3026 network interfaces for predictable names, and a combination
3027 of these policies is now the default. Please see this wiki
3028 document for details:
3030 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
3032 * Auke Kok's bootchart implementation has been added to the
3033 systemd tree. It is an optional component that can graph the
3034 boot in quite some detail. It is one of the best bootchart
3035 implementations around and minimal in its code and
3038 * nss-myhostname has been integrated into the systemd source
3039 tree. nss-myhostname guarantees that the local hostname
3040 always stays resolvable via NSS. It has been a weak
3041 requirement of systemd-hostnamed since a long time, and
3042 since its code is actually trivial we decided to just
3043 include it in systemd's source tree. It can be turned off
3044 with a configure switch.
3046 * The read-ahead logic is now capable of properly detecting
3047 whether a btrfs file system is on SSD or rotating media, in
3048 order to optimize the read-ahead scheme. Previously, it was
3049 only capable of detecting this on traditional file systems
3052 * In udev, additional device properties are now read from the
3053 IAB in addition to the OUI database. Also, Bluetooth company
3054 identities are attached to the devices as well.
3056 * In service files %U may be used as specifier that is
3057 replaced by the configured user name of the service.
3059 * nspawn may now be invoked without a controlling TTY. This
3060 makes it suitable for invocation as its own service. This
3061 may be used to set up a simple containerized server system
3062 using only core OS tools.
3064 * systemd and nspawn can now accept socket file descriptors
3065 when they are started for socket activation. This enables
3066 implementation of socket activated nspawn
3067 containers. i.e. think about autospawning an entire OS image
3068 when the first SSH or HTTP connection is received. We expect
3069 that similar functionality will also be added to libvirt-lxc
3072 * journalctl will now suppress ANSI color codes when
3073 presenting log data.
3075 * systemctl will no longer show control group information for
3076 a unit if a the control group is empty anyway.
3078 * logind can now automatically suspend/hibernate/shutdown the
3081 * /etc/machine-info and hostnamed now also expose the chassis
3082 type of the system. This can be used to determine whether
3083 the local system is a laptop, desktop, handset or
3084 tablet. This information may either be configured by the
3085 user/vendor or is automatically determined from ACPI and DMI
3086 information if possible.
3088 * A number of PolicyKit actions are now bound together with
3089 "imply" rules. This should simplify creating UIs because
3090 many actions will now authenticate similar ones as well.
3092 * Unit files learnt a new condition ConditionACPower= which
3093 may be used to conditionalize a unit depending on whether an
3094 AC power source is connected or not, of whether the system
3095 is running on battery power.
3097 * systemctl gained a new "is-failed" verb that may be used in
3098 shell scripts and suchlike to check whether a specific unit
3099 is in the "failed" state.
3101 * The EnvironmentFile= setting in unit files now supports file
3102 globbing, and can hence be used to easily read a number of
3103 environment files at once.
3105 * systemd will no longer detect and recognize specific
3106 distributions. All distribution-specific #ifdeffery has been
3107 removed, systemd is now fully generic and
3108 distribution-agnostic. Effectively, not too much is lost as
3109 a lot of the code is still accessible via explicit configure
3110 switches. However, support for some distribution specific
3111 legacy configuration file formats has been dropped. We
3112 recommend distributions to simply adopt the configuration
3113 files everybody else uses now and convert the old
3114 configuration from packaging scripts. Most distributions
3115 already did that. If that's not possible or desirable,
3116 distributions are welcome to forward port the specific
3117 pieces of code locally from the git history.
3119 * When logging a message about a unit systemd will now always
3120 log the unit name in the message meta data.
3122 * localectl will now also discover system locale data that is
3123 not stored in locale archives, but directly unpacked.
3125 * logind will no longer unconditionally use framebuffer
3126 devices as seat masters, i.e. as devices that are required
3127 to be existing before a seat is considered preset. Instead,
3128 it will now look for all devices that are tagged as
3129 "seat-master" in udev. By default framebuffer devices will
3130 be marked as such, but depending on local systems other
3131 devices might be marked as well. This may be used to
3132 integrate graphics cards using closed source drivers (such
3133 as NVidia ones) more nicely into logind. Note however, that
3134 we recommend using the open source NVidia drivers instead,
3135 and no udev rules for the closed-source drivers will be
3136 shipped from us upstream.
3138 Contributions from: Adam Williamson, Alessandro Crismani, Auke
3139 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
3140 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
3141 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
3142 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
3143 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
3144 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
3145 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
3146 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
3147 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
3148 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
3149 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
3154 * udev gained support for loading additional device properties
3155 from an indexed database that is keyed by vendor/product IDs
3156 and similar device identifiers. For the beginning this
3157 "hwdb" is populated with data from the well-known PCI and
3158 USB database, but also includes PNP, ACPI and OID data. In
3159 the longer run this indexed database shall grow into
3160 becoming the one central database for non-essential
3161 userspace device metadata. Previously, data from the PCI/USB
3162 database was only attached to select devices, since the
3163 lookup was a relatively expensive operation due to O(n) time
3164 complexity (with n being the number of entries in the
3165 database). Since this is now O(1), we decided to add in this
3166 data for all devices where this is available, by
3167 default. Note that the indexed database needs to be rebuilt
3168 when new data files are installed. To achieve this you need
3169 to update your packaging scripts to invoke "udevadm hwdb
3170 --update" after installation of hwdb data files. For
3171 RPM-based distributions we introduced the new
3172 %udev_hwdb_update macro for this purpose.
3174 * The Journal gained support for the "Message Catalog", an
3175 indexed database to link up additional information with
3176 journal entries. For further details please check:
3178 http://www.freedesktop.org/wiki/Software/systemd/catalog
3180 The indexed message catalog database also needs to be
3181 rebuilt after installation of message catalog files. Use
3182 "journalctl --update-catalog" for this. For RPM-based
3183 distributions we introduced the %journal_catalog_update
3184 macro for this purpose.
3186 * The Python Journal bindings gained support for the standard
3187 Python logging framework.
3189 * The Journal API gained new functions for checking whether
3190 the underlying file system of a journal file is capable of
3191 properly reporting file change notifications, or whether
3192 applications that want to reflect journal changes "live"
3193 need to recheck journal files continuously in appropriate
3196 * It is now possible to set the "age" field for tmpfiles
3197 entries to 0, indicating that files matching this entry
3198 shall always be removed when the directories are cleaned up.
3200 * coredumpctl gained a new "gdb" verb which invokes gdb
3201 right-away on the selected coredump.
3203 * There's now support for "hybrid sleep" on kernels that
3204 support this, in addition to "suspend" and "hibernate". Use
3205 "systemctl hybrid-sleep" to make use of this.
3207 * logind's HandleSuspendKey= setting (and related settings)
3208 now gained support for a new "lock" setting to simply
3209 request the screen lock on all local sessions, instead of
3210 actually executing a suspend or hibernation.
3212 * systemd will now mount the EFI variables file system by
3215 * Socket units now gained support for configuration of the
3216 SMACK security label.
3218 * timedatectl will now output the time of the last and next
3219 daylight saving change.
3221 * We dropped support for various legacy and distro-specific
3222 concepts, such as insserv, early-boot SysV services
3223 (i.e. those for non-standard runlevels such as 'b' or 'S')
3224 or ArchLinux /etc/rc.conf support. We recommend the
3225 distributions who still need support this to either continue
3226 to maintain the necessary patches downstream, or find a
3227 different solution. (Talk to us if you have questions!)
3229 * Various systemd components will now bypass PolicyKit checks
3230 for root and otherwise handle properly if PolicyKit is not
3231 found to be around. This should fix most issues for
3232 PolicyKit-less systems. Quite frankly this should have been
3233 this way since day one. It is absolutely our intention to
3234 make systemd work fine on PolicyKit-less systems, and we
3235 consider it a bug if something does not work as it should if
3236 PolicyKit is not around.
3238 * For embedded systems it is now possible to build udev and
3239 systemd without blkid and/or kmod support.
3241 * "systemctl switch-root" is now capable of switching root
3242 more than once. I.e. in addition to transitions from the
3243 initrd to the host OS it is now possible to transition to
3244 further OS images from the host. This is useful to implement
3245 offline updating tools.
3247 * Various other additions have been made to the RPM macros
3248 shipped with systemd. Use %udev_rules_update() after
3249 installing new udev rules files. %_udevhwdbdir,
3250 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
3251 %_sysctldir are now available which resolve to the right
3252 directories for packages to place various data files in.
3254 * journalctl gained the new --full switch (in addition to
3255 --all, to disable ellipsation for long messages.
3257 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
3258 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
3259 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
3260 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
3261 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
3262 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
3263 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
3264 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
3265 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
3269 * journalctl gained new --since= and --until= switches to
3270 filter by time. It also now supports nice filtering for
3271 units via --unit=/-u.
3273 * Type=oneshot services may use ExecReload= and do the
3276 * The journal daemon now supports time-based rotation and
3277 vacuuming, in addition to the usual disk-space based
3280 * The journal will now index the available field values for
3281 each field name. This enables clients to show pretty drop
3282 downs of available match values when filtering. The bash
3283 completion of journalctl has been updated
3284 accordingly. journalctl gained a new switch -F to list all
3285 values a certain field takes in the journal database.
3287 * More service events are now written as structured messages
3288 to the journal, and made recognizable via message IDs.
3290 * The timedated, localed and hostnamed mini-services which
3291 previously only provided support for changing time, locale
3292 and hostname settings from graphical DEs such as GNOME now
3293 also have a minimal (but very useful) text-based client
3294 utility each. This is probably the nicest way to changing
3295 these settings from the command line now, especially since
3296 it lists available options and is fully integrated with bash
3299 * There's now a new tool "systemd-coredumpctl" to list and
3300 extract coredumps from the journal.
3302 * We now install a README each in /var/log/ and
3303 /etc/rc.d/init.d explaining where the system logs and init
3304 scripts went. This hopefully should help folks who go to
3305 that dirs and look into the otherwise now empty void and
3306 scratch their heads.
3308 * When user-services are invoked (by systemd --user) the
3309 $MANAGERPID env var is set to the PID of systemd.
3311 * SIGRTMIN+24 when sent to a --user instance will now result
3312 in immediate termination of systemd.
3314 * gatewayd received numerous feature additions such as a
3315 "follow" mode, for live syncing and filtering.
3317 * browse.html now allows filtering and showing detailed
3318 information on specific entries. Keyboard navigation and
3319 mouse screen support has been added.
3321 * gatewayd/journalctl now supports HTML5/JSON
3322 Server-Sent-Events as output.
3324 * The SysV init script compatibility logic will now
3325 heuristically determine whether a script supports the
3326 "reload" verb, and only then make this available as
3329 * "systemctl status --follow" has been removed, use "journalctl
3332 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
3333 have been removed since they are hardly useful to be
3336 * And I'd like to take the opportunity to specifically mention
3337 Zbigniew for his great contributions. Zbigniew, you rock!
3339 Contributions from: Andrew Eikum, Christian Hesse, Colin
3340 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
3341 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
3342 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
3343 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
3344 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
3345 Jędrzejewski-Szmek, Сковорода Никита Андреевич
3349 * If /etc/vconsole.conf is non-existent or empty we will no
3350 longer load any console font or key map at boot by
3351 default. Instead the kernel defaults will be left
3352 intact. This is definitely the right thing to do, as no
3353 configuration should mean no configuration, and hard-coding
3354 font names that are different on all archs is probably a bad
3355 idea. Also, the kernel default key map and font should be
3356 good enough for most cases anyway, and mostly identical to
3357 the userspace fonts/key maps we previously overloaded them
3358 with. If distributions want to continue to default to a
3359 non-kernel font or key map they should ship a default
3360 /etc/vconsole.conf with the appropriate contents.
3362 Contributions from: Colin Walters, Daniel J Walsh, Dave
3363 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
3364 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
3368 * journalctl gained a new --cursor= switch to show entries
3369 starting from the specified location in the journal.
3371 * We now enforce a size limit on journal entry fields exported
3372 with "-o json" in journalctl. Fields larger than 4K will be
3373 assigned null. This can be turned off with --all.
3375 * An (optional) journal gateway daemon is now available as
3376 "systemd-journal-gatewayd.service". This service provides
3377 access to the journal via HTTP and JSON. This functionality
3378 will be used to implement live log synchronization in both
3379 pull and push modes, but has various other users too, such
3380 as easy log access for debugging of embedded devices. Right
3381 now it is already useful to retrieve the journal via HTTP:
3383 # systemctl start systemd-journal-gatewayd.service
3384 # wget http://localhost:19531/entries
3386 This will download the journal contents in a
3387 /var/log/messages compatible format. The same as JSON:
3389 # curl -H"Accept: application/json" http://localhost:19531/entries
3391 This service is also accessible via a web browser where a
3392 single static HTML5 app is served that uses the JSON logic
3393 to enable the user to do some basic browsing of the
3394 journal. This will be extended later on. Here's an example
3395 screenshot of this app in its current state:
3397 http://0pointer.de/public/journal-gatewayd
3399 Contributions from: Kay Sievers, Lennart Poettering, Robert
3400 Milasan, Tom Gundersen
3404 * The bash completion logic is now available for journalctl
3407 * We do not mount the "cpuset" controller anymore together with
3408 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
3409 started if no parameters are assigned to it. "cpuset" hence
3410 broke code that assumed it it could create "cpu" groups and
3413 * journalctl -f will now subscribe to terminal size changes,
3414 and line break accordingly.
3416 Contributions from: Dave Reisner, Kay Sievers, Lennart
3417 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
3421 * nspawn will now create a symlink /etc/localtime in the
3422 container environment, copying the host's timezone
3423 setting. Previously this has been done via a bind mount, but
3424 since symlinks cannot be bind mounted this has now been
3425 changed to create/update the appropriate symlink.
3427 * journalctl -n's line number argument is now optional, and
3428 will default to 10 if omitted.
3430 * journald will now log the maximum size the journal files may
3431 take up on disk. This is particularly useful if the default
3432 built-in logic of determining this parameter from the file
3433 system size is used. Use "systemctl status
3434 systemd-journald.service" to see this information.
3436 * The multi-seat X wrapper tool has been stripped down. As X
3437 is now capable of enumerating graphics devices via udev in a
3438 seat-aware way the wrapper is not strictly necessary
3439 anymore. A stripped down temporary stop-gap is still shipped
3440 until the upstream display managers have been updated to
3441 fully support the new X logic. Expect this wrapper to be
3442 removed entirely in one of the next releases.
3444 * HandleSleepKey= in logind.conf has been split up into
3445 HandleSuspendKey= and HandleHibernateKey=. The old setting
3446 is not available anymore. X11 and the kernel are
3447 distuingishing between these keys and we should too. This
3448 also means the inhibition lock for these keys has been split
3451 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
3452 Poettering, Lukas Nykryn, Václav Pavlín
3456 * Whenever a unit changes state we will now log this to the
3457 journal and show along the unit's own log output in
3460 * ConditionPathIsMountPoint= can now properly detect bind
3461 mount points too. (Previously, a bind mount of one file
3462 system to another place in the same file system could not be
3463 detected as mount, since they shared struct stat's st_dev
3466 * We will now mount the cgroup controllers cpu, cpuacct,
3467 cpuset and the controllers net_cls, net_prio together by
3470 * nspawn containers will now have a virtualized boot
3471 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
3472 over with a randomized ID at container initialization). This
3473 has the effect of making "journalctl -b" do the right thing
3476 * The JSON output journal serialization has been updated not
3477 to generate "endless" list objects anymore, but rather one
3478 JSON object per line. This is more in line how most JSON
3479 parsers expect JSON objects. The new output mode
3480 "json-pretty" has been added to provide similar output, but
3481 neatly aligned for readability by humans.
3483 * We dropped all explicit sync() invocations in the shutdown
3484 code. The kernel does this implicitly anyway in the kernel
3485 reboot() syscall. halt(8)'s -n option is now a compatibility
3488 * We now support virtualized reboot() in containers, as
3489 supported by newer kernels. We will fall back to exit() if
3490 CAP_SYS_REBOOT is not available to the container. Also,
3491 nspawn makes use of this now and will actually reboot the
3492 container if the containerized OS asks for that.
3494 * journalctl will only show local log output by default
3495 now. Use --merge (-m) to show remote log output, too.
3497 * libsystemd-journal gained the new sd_journal_get_usage()
3498 call to determine the current disk usage of all journal
3499 files. This is exposed in the new "journalctl --disk-usage"
3502 * journald gained a new configuration setting SplitMode= in
3503 journald.conf which may be used to control how user journals
3504 are split off. See journald.conf(5) for details.
3506 * A new condition type ConditionFileNotEmpty= has been added.
3508 * tmpfiles' "w" lines now support file globbing, to write
3509 multiple files at once.
3511 * We added Python bindings for the journal submission
3512 APIs. More Python APIs for a number of selected APIs will
3513 likely follow. Note that we intend to add native bindings
3514 only for the Python language, as we consider it common
3515 enough to deserve bindings shipped within systemd. There are
3516 various projects outside of systemd that provide bindings
3517 for languages such as PHP or Lua.
3519 * Many conditions will now resolve specifiers such as %i. In
3520 addition, PathChanged= and related directives of .path units
3521 now support specifiers as well.
3523 * There's now a new RPM macro definition for the system preset
3526 * journald will now warn if it ca not forward a message to the
3527 syslog daemon because its socket is full.
3529 * timedated will no longer write or process /etc/timezone,
3530 except on Debian. As we do not support late mounted /usr
3531 anymore /etc/localtime always being a symlink is now safe,
3532 and hence the information in /etc/timezone is not necessary
3535 * logind will now always reserve one VT for a text getty (VT6
3536 by default). Previously if more than 6 X sessions where
3537 started they took up all the VTs with auto-spawned gettys,
3538 so that no text gettys were available anymore.
3540 * udev will now automatically inform the btrfs kernel logic
3541 about btrfs RAID components showing up. This should make
3542 simple hotplug based btrfs RAID assembly work.
3544 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
3545 (but not for its children which will stay at the kernel
3546 default). This should allow setups with a lot more listening
3549 * systemd will now always pass the configured timezone to the
3550 kernel at boot. timedated will do the same when the timezone
3553 * logind's inhibition logic has been updated. By default,
3554 logind will now handle the lid switch, the power and sleep
3555 keys all the time, even in graphical sessions. If DEs want
3556 to handle these events on their own they should take the new
3557 handle-power-key, handle-sleep-key and handle-lid-switch
3558 inhibitors during their runtime. A simple way to achieve
3559 that is to invoke the DE wrapped in an invocation of:
3561 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
3563 * Access to unit operations is now checked via SELinux taking
3564 the unit file label and client process label into account.
3566 * systemd will now notify the administrator in the journal
3567 when he over-mounts a non-empty directory.
3569 * There are new specifiers that are resolved in unit files,
3570 for the host name (%H), the machine ID (%m) and the boot ID
3573 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
3574 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
3575 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
3576 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
3577 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
3578 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
3579 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
3583 * Support for reading structured kernel messages from
3584 /dev/kmsg has now been added and is enabled by default.
3586 * Support for reading kernel messages from /proc/kmsg has now
3587 been removed. If you want kernel messages in the journal
3588 make sure to run a recent kernel (>= 3.5) that supports
3589 reading structured messages from /dev/kmsg (see
3590 above). /proc/kmsg is now exclusive property of classic
3591 syslog daemons again.
3593 * The libudev API gained the new
3594 udev_device_new_from_device_id() call.
3596 * The logic for file system namespace (ReadOnlyDirectory=,
3597 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
3598 require pivot_root() anymore. This means fewer temporary
3599 directories are created below /tmp for this feature.
3601 * nspawn containers will now see and receive all submounts
3602 made on the host OS below the root file system of the
3605 * Forward Secure Sealing is now supported for Journal files,
3606 which provide cryptographical sealing of journal files so
3607 that attackers cannot alter log history anymore without this
3608 being detectable. Lennart will soon post a blog story about
3609 this explaining it in more detail.
3611 * There are two new service settings RestartPreventExitStatus=
3612 and SuccessExitStatus= which allow configuration of exit
3613 status (exit code or signal) which will be excepted from the
3614 restart logic, resp. consider successful.
3616 * journalctl gained the new --verify switch that can be used
3617 to check the integrity of the structure of journal files and
3618 (if Forward Secure Sealing is enabled) the contents of
3621 * nspawn containers will now be run with /dev/stdin, /dev/fd/
3622 and similar symlinks pre-created. This makes running shells
3623 as container init process a lot more fun.
3625 * The fstab support can now handle PARTUUID= and PARTLABEL=
3628 * A new ConditionHost= condition has been added to match
3629 against the hostname (with globs) and machine ID. This is
3630 useful for clusters where a single OS image is used to
3631 provision a large number of hosts which shall run slightly
3632 different sets of services.
3634 * Services which hit the restart limit will now be placed in a
3637 Contributions from: Bertram Poettering, Dave Reisner, Huang
3638 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
3639 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
3643 * When running in --user mode systemd will now become a
3644 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
3645 tree a lot more organized.
3647 * A new PartOf= unit dependency type has been introduced that
3648 may be used to group services in a natural way.
3650 * "systemctl enable" may now be used to enable instances of
3653 * journalctl now prints error log levels in red, and
3654 warning/notice log levels in bright white. It also supports
3655 filtering by log level now.
3657 * cgtop gained a new -n switch (similar to top), to configure
3658 the maximum number of iterations to run for. It also gained
3659 -b, to run in batch mode (accepting no input).
3661 * The suffix ".service" may now be omitted on most systemctl
3662 command lines involving service unit names.
3664 * There's a new bus call in logind to lock all sessions, as
3665 well as a loginctl verb for it "lock-sessions".
3667 * libsystemd-logind.so gained a new call sd_journal_perror()
3668 that works similar to libc perror() but logs to the journal
3669 and encodes structured information about the error number.
3671 * /etc/crypttab entries now understand the new keyfile-size=
3674 * shutdown(8) now can send a (configurable) wall message when
3675 a shutdown is cancelled.
3677 * The mount propagation mode for the root file system will now
3678 default to "shared", which is useful to make containers work
3679 nicely out-of-the-box so that they receive new mounts from
3680 the host. This can be undone locally by running "mount
3681 --make-rprivate /" if needed.
3683 * The prefdm.service file has been removed. Distributions
3684 should maintain this unit downstream if they intend to keep
3685 it around. However, we recommend writing normal unit files
3686 for display managers instead.
3688 * Since systemd is a crucial part of the OS we will now
3689 default to a number of compiler switches that improve
3690 security (hardening) such as read-only relocations, stack
3691 protection, and suchlike.
3693 * The TimeoutSec= setting for services is now split into
3694 TimeoutStartSec= and TimeoutStopSec= to allow configuration
3695 of individual time outs for the start and the stop phase of
3698 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
3699 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
3700 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
3701 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
3702 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
3703 Gundersen, Zbigniew Jędrzejewski-Szmek
3707 * The journal and id128 C APIs are now fully documented as man
3710 * Extra safety checks have been added when transitioning from
3711 the initial RAM disk to the main system to avoid accidental
3714 * /etc/crypttab entries now understand the new keyfile-offset=
3717 * systemctl -t can now be used to filter by unit load state.
3719 * The journal C API gained the new sd_journal_wait() call to
3720 make writing synchronous journal clients easier.
3722 * journalctl gained the new -D switch to show journals from a
3725 * journalctl now displays a special marker between log
3726 messages of two different boots.
3728 * The journal is now explicitly flushed to /var via a service
3729 systemd-journal-flush.service, rather than implicitly simply
3730 by seeing /var/log/journal to be writable.
3732 * journalctl (and the journal C APIs) can now match for much
3733 more complex expressions, with alternatives and
3736 * When transitioning from the initial RAM disk to the main
3737 system we will now kill all processes in a killing spree to
3738 ensure no processes stay around by accident.
3740 * Three new specifiers may be used in unit files: %u, %h, %s
3741 resolve to the user name, user home directory resp. user
3742 shell. This is useful for running systemd user instances.
3744 * We now automatically rotate journal files if their data
3745 object hash table gets a fill level > 75%. We also size the
3746 hash table based on the configured maximum file size. This
3747 together should lower hash collisions drastically and thus
3748 speed things up a bit.
3750 * journalctl gained the new "--header" switch to introspect
3751 header data of journal files.
3753 * A new setting SystemCallFilters= has been added to services
3754 which may be used to apply blacklists or whitelists to
3755 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
3757 * nspawn gained a new --link-journal= switch (and quicker: -j)
3758 to link the container journal with the host. This makes it
3759 very easy to centralize log viewing on the host for all
3760 guests while still keeping the journal files separated.
3762 * Many bugfixes and optimizations
3764 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
3765 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
3766 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
3771 * Several tools now understand kernel command line arguments,
3772 which are only read when run in an initial RAM disk. They
3773 usually follow closely their normal counterparts, but are
3776 * There's a new tool to analyze the readahead files that are
3777 automatically generated at boot. Use:
3779 /usr/lib/systemd/systemd-readahead analyze /.readahead
3781 * We now provide an early debug shell on tty9 if this enabled. Use:
3783 systemctl enable debug-shell.service
3785 * All plymouth related units have been moved into the Plymouth
3786 package. Please make sure to upgrade your Plymouth version
3789 * systemd-tmpfiles now supports getting passed the basename of
3790 a configuration file only, in which case it will look for it
3791 in all appropriate directories automatically.
3793 * udevadm info now takes a /dev or /sys path as argument, and
3794 does the right thing. Example:
3796 udevadm info /dev/sda
3797 udevadm info /sys/class/block/sda
3799 * systemctl now prints a warning if a unit is stopped but a
3800 unit that might trigger it continues to run. Example: a
3801 service is stopped but the socket that activates it is left
3804 * "systemctl status" will now mention if the log output was
3805 shortened due to rotation since a service has been started.
3807 * The journal API now exposes functions to determine the
3808 "cutoff" times due to rotation.
3810 * journald now understands SIGUSR1 and SIGUSR2 for triggering
3811 immediately flushing of runtime logs to /var if possible,
3812 resp. for triggering immediate rotation of the journal
3815 * It is now considered an error if a service is attempted to
3816 be stopped that is not loaded.
3818 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
3820 * systemd-analyze now supports Python 3
3822 * tmpfiles now supports cleaning up directories via aging
3823 where the first level dirs are always kept around but
3824 directories beneath it automatically aged. This is enabled
3825 by prefixing the age field with '~'.
3827 * Seat objects now expose CanGraphical, CanTTY properties
3828 which is required to deal with very fast bootups where the
3829 display manager might be running before the graphics drivers
3830 completed initialization.
3832 * Seat objects now expose a State property.
3834 * We now include RPM macros for service enabling/disabling
3835 based on the preset logic. We recommend RPM based
3836 distributions to make use of these macros if possible. This
3837 makes it simpler to reuse RPM spec files across
3840 * We now make sure that the collected systemd unit name is
3841 always valid when services log to the journal via
3844 * There's a new man page kernel-command-line(7) detailing all
3845 command line options we understand.
3847 * The fstab generator may now be disabled at boot by passing
3848 fstab=0 on the kernel command line.
3850 * A new kernel command line option modules-load= is now understood
3851 to load a specific kernel module statically, early at boot.
3853 * Unit names specified on the systemctl command line are now
3854 automatically escaped as needed. Also, if file system or
3855 device paths are specified they are automatically turned
3856 into the appropriate mount or device unit names. Example:
3858 systemctl status /home
3859 systemctl status /dev/sda
3861 * The SysVConsole= configuration option has been removed from
3862 system.conf parsing.
3864 * The SysV search path is no longer exported on the D-Bus
3867 * The Names= option is been removed from unit file parsing.
3869 * There's a new man page bootup(7) detailing the boot process.
3871 * Every unit and every generator we ship with systemd now
3872 comes with full documentation. The self-explanatory boot is
3875 * A couple of services gained "systemd-" prefixes in their
3876 name if they wrap systemd code, rather than only external
3877 code. Among them fsck@.service which is now
3878 systemd-fsck@.service.
3880 * The HaveWatchdog property has been removed from the D-Bus
3883 * systemd.confirm_spawn= on the kernel command line should now
3886 * There's a new man page crypttab(5) which details all options
3887 we actually understand.
3889 * systemd-nspawn gained a new --capability= switch to pass
3890 additional capabilities to the container.
3892 * timedated will now read known NTP implementation unit names
3893 from /usr/lib/systemd/ntp-units.d/*.list,
3894 systemd-timedated-ntp.target has been removed.
3896 * journalctl gained a new switch "-b" that lists log data of
3897 the current boot only.
3899 * The notify socket is in the abstract namespace again, in
3900 order to support daemons which chroot() at start-up.
3902 * There is a new Storage= configuration option for journald
3903 which allows configuration of where log data should go. This
3904 also provides a way to disable journal logging entirely, so
3905 that data collected is only forwarded to the console, the
3906 kernel log buffer or another syslog implementation.
3908 * Many bugfixes and optimizations
3910 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
3911 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
3912 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
3913 Shawn Landden, Tom Gundersen
3917 * "systemctl help <unit>" now shows the man page if one is
3920 * Several new man pages have been added.
3922 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
3923 MaxLevelConsole= can now be specified in
3924 journald.conf. These options allow reducing the amount of
3925 data stored on disk or forwarded by the log level.
3927 * TimerSlackNSec= can now be specified in system.conf for
3928 PID1. This allows system-wide power savings.
3930 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
3931 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
3936 * logind is now capable of (optionally) handling power and
3937 sleep keys as well as the lid switch.
3939 * journalctl now understands the syntax "journalctl
3940 /usr/bin/avahi-daemon" to get all log output of a specific
3943 * CapabilityBoundingSet= in system.conf now also influences
3944 the capability bound set of usermode helpers of the kernel.
3946 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
3947 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
3948 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
3949 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
3953 * Note that we skipped 139 releases here in order to set the
3954 new version to something that is greater than both udev's
3955 and systemd's most recent version number.
3957 * udev: all udev sources are merged into the systemd source tree now.
3958 All future udev development will happen in the systemd tree. It
3959 is still fully supported to use the udev daemon and tools without
3960 systemd running, like in initramfs or other init systems. Building
3961 udev though, will require the *build* of the systemd tree, but
3962 udev can be properly *run* without systemd.
3964 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
3965 should be used to create dead device nodes as workarounds for broken
3968 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
3969 no longer supported. udev_monitor_new_from_netlink() needs to be
3970 used to subscribe to events.
3972 * udev: when udevd is started by systemd, processes which are left
3973 behind by forking them off of udev rules, are unconditionally cleaned
3974 up and killed now after the event handling has finished. Services or
3975 daemons must be started as systemd services. Services can be
3976 pulled-in by udev to get started, but they can no longer be directly
3977 forked by udev rules.
3979 * udev: the daemon binary is called systemd-udevd now and installed
3980 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
3981 to adapt to that, create symlink, or rename the binary after building
3984 * libudev no longer provides these symbols:
3985 udev_monitor_from_socket()
3986 udev_queue_get_failed_list_entry()
3987 udev_get_{dev,sys,run}_path()
3988 The versions number was bumped and symbol versioning introduced.
3990 * systemd-loginctl and systemd-journalctl have been renamed
3991 to loginctl and journalctl to match systemctl.
3993 * The config files: /etc/systemd/systemd-logind.conf and
3994 /etc/systemd/systemd-journald.conf have been renamed to
3995 logind.conf and journald.conf. Package updates should rename
3996 the files to the new names on upgrade.
3998 * For almost all files the license is now LGPL2.1+, changed
3999 from the previous GPL2.0+. Exceptions are some minor stuff
4000 of udev (which will be changed to LGPL2.1 eventually, too),
4001 and the MIT licensed sd-daemon.[ch] library that is suitable
4002 to be used as drop-in files.
4004 * systemd and logind now handle system sleep states, in
4005 particular suspending and hibernating.
4007 * logind now implements a sleep/shutdown/idle inhibiting logic
4008 suitable for a variety of uses. Soonishly Lennart will blog
4009 about this in more detail.
4011 * var-run.mount and var-lock.mount are no longer provided
4012 (which prevously bind mounted these directories to their new
4013 places). Distributions which have not converted these
4014 directories to symlinks should consider stealing these files
4015 from git history and add them downstream.
4017 * We introduced the Documentation= field for units and added
4018 this to all our shipped units. This is useful to make it
4019 easier to explore the boot and the purpose of the various
4022 * All smaller setup units (such as
4023 systemd-vconsole-setup.service) now detect properly if they
4024 are run in a container and are skipped when
4025 appropriate. This guarantees an entirely noise-free boot in
4026 Linux container environments such as systemd-nspawn.
4028 * A framework for implementing offline system updates is now
4029 integrated, for details see:
4030 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
4032 * A new service type Type=idle is available now which helps us
4033 avoiding ugly interleaving of getty output and boot status
4036 * There's now a system-wide CapabilityBoundingSet= option to
4037 globally reduce the set of capabilities for the
4038 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
4039 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
4040 even CAP_NET_ADMIN system-wide for secure systems.
4042 * There are now system-wide DefaultLimitXXX= options to
4043 globally change the defaults of the various resource limits
4044 for all units started by PID 1.
4046 * Harald Hoyer's systemd test suite has been integrated into
4047 systemd which allows easy testing of systemd builds in qemu
4048 and nspawn. (This is really awesome! Ask us for details!)
4050 * The fstab parser is now implemented as generator, not inside
4053 * systemctl will now warn you if .mount units generated from
4054 /etc/fstab are out of date due to changes in fstab that
4055 have not been read by systemd yet.
4057 * systemd is now suitable for usage in initrds. Dracut has
4058 already been updated to make use of this. With this in place
4059 initrds get a slight bit faster but primarily are much
4060 easier to introspect and debug since "systemctl status" in
4061 the host system can be used to introspect initrd services,
4062 and the journal from the initrd is kept around too.
4064 * systemd-delta has been added, a tool to explore differences
4065 between user/admin configuration and vendor defaults.
4067 * PrivateTmp= now affects both /tmp and /var/tmp.
4069 * Boot time status messages are now much prettier and feature
4070 proper english language. Booting up systemd has never been
4073 * Read-ahead pack files now include the inode number of all
4074 files to pre-cache. When the inode changes the pre-caching
4075 is not attempted. This should be nicer to deal with updated
4076 packages which might result in changes of read-ahead
4079 * We now temporaritly lower the kernel's read_ahead_kb variable
4080 when collecting read-ahead data to ensure the kernel's
4081 built-in read-ahead does not add noise to our measurements
4082 of necessary blocks to pre-cache.
4084 * There's now RequiresMountsFor= to add automatic dependencies
4085 for all mounts necessary for a specific file system path.
4087 * MountAuto= and SwapAuto= have been removed from
4088 system.conf. Mounting file systems at boot has to take place
4091 * nspawn now learned a new switch --uuid= to set the machine
4092 ID on the command line.
4094 * nspawn now learned the -b switch to automatically search
4097 * vt102 is now the default TERM for serial TTYs, upgraded from
4100 * systemd-logind now works on VT-less systems.
4102 * The build tree has been reorganized. The individual
4103 components now have directories of their own.
4105 * A new condition type ConditionPathIsReadWrite= is now available.
4107 * nspawn learned the new -C switch to create cgroups for the
4108 container in other hierarchies.
4110 * We now have support for hardware watchdogs, configurable in
4113 * The scheduled shutdown logic now has a public API.
4115 * We now mount /tmp as tmpfs by default, but this can be
4116 masked and /etc/fstab can override it.
4118 * Since udisks does not make use of /media anymore we are not
4119 mounting a tmpfs on it anymore.
4121 * journalctl gained a new --local switch to only interleave
4122 locally generated journal files.
4124 * We can now load the IMA policy at boot automatically.
4126 * The GTK tools have been split off into a systemd-ui.
4128 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
4129 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
4130 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
4131 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
4132 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
4133 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
4134 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
4135 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
4136 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
4141 * This is mostly a bugfix release
4143 * Support optional initialization of the machine ID from the
4144 KVM or container configured UUID.
4146 * Support immediate reboots with "systemctl reboot -ff"
4148 * Show /etc/os-release data in systemd-analyze output
4150 * Many bugfixes for the journal, including endianness fixes and
4151 ensuring that disk space enforcement works
4153 * sd-login.h is C++ comptaible again
4155 * Extend the /etc/os-release format on request of the Debian
4158 * We now refuse non-UTF8 strings used in various configuration
4159 and unit files. This is done to ensure we do not pass invalid
4160 data over D-Bus or expose it elsewhere.
4162 * Register Mimo USB Screens as suitable for automatic seat
4165 * Read SELinux client context from journal clients in a race
4168 * Reorder configuration file lookup order. /etc now always
4169 overrides /run in order to allow the administrator to always
4170 and unconditionally override vendor supplied or
4171 automatically generated data.
4173 * The various user visible bits of the journal now have man
4174 pages. We still lack man pages for the journal API calls
4177 * We now ship all man pages in HTML format again in the
4180 Contributions from: Dave Reisner, Dirk Eibach, Frederic
4181 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
4182 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
4187 * This is mostly a bugfix release
4189 * systems lacking /etc/os-release are no longer supported.
4191 * Various functionality updates to libsystemd-login.so
4193 * Track class of PAM logins to distuingish greeters from
4196 Contributions from: Kay Sievers, Lennart Poettering, Michael
4201 * This is an important bugfix release for v41.
4203 * Building man pages is now optional which should be useful
4204 for those building systemd from git but unwilling to install
4207 * Watchdog support for supervising services is now usable. In
4208 a future release support for hardware watchdogs
4209 (i.e. /dev/watchdog) will be added building on this.
4211 * Service start rate limiting is now configurable and can be
4212 turned off per service. When a start rate limit is hit a
4213 reboot can automatically be triggered.
4215 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
4217 Contributions from: Benjamin Franzke, Bill Nottingham,
4218 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
4219 Schmidt, Michał Górny, Piotr Drąg
4223 * The systemd binary is installed /usr/lib/systemd/systemd now;
4224 An existing /sbin/init symlink needs to be adapted with the
4227 * The code that loads kernel modules has been ported to invoke
4228 libkmod directly, instead of modprobe. This means we do not
4229 support systems with module-init-tools anymore.
4231 * Watchdog support is now already useful, but still not
4234 * A new kernel command line option systemd.setenv= is
4235 understood to set system wide environment variables
4236 dynamically at boot.
4238 * We now limit the set of capabilities of systemd-journald.
4240 * We now set SIGPIPE to ignore by default, since it only is
4241 useful in shell pipelines, and has little use in general
4242 code. This can be disabled with IgnoreSIPIPE=no in unit
4245 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
4246 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
4251 * This is mostly a bugfix release
4253 * We now expose the reason why a service failed in the
4254 "Result" D-Bus property.
4256 * Rudimentary service watchdog support (will be completed over
4257 the next few releases.)
4259 * When systemd forks off in order execute some service we will
4260 now immediately changes its argv[0] to reflect which process
4261 it will execute. This is useful to minimize the time window
4262 with a generic argv[0], which makes bootcharts more useful
4264 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
4265 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
4266 Mike Kazantsev, Ray Strode
4270 * This is mostly a test release, but incorporates many
4273 * New systemd-cgtop tool to show control groups by their
4276 * Linking against libacl for ACLs is optional again. If
4277 disabled, support tracking device access for active logins
4278 goes becomes unavailable, and so does access to the user
4279 journals by the respective users.
4281 * If a group "adm" exists, journal files are automatically
4282 owned by them, thus allow members of this group full access
4283 to the system journal as well as all user journals.
4285 * The journal now stores the SELinux context of the logging
4286 client for all entries.
4288 * Add C++ inclusion guards to all public headers
4290 * New output mode "cat" in the journal to print only text
4291 messages, without any meta data like date or time.
4293 * Include tiny X server wrapper as a temporary stop-gap to
4294 teach XOrg udev display enumeration. This is used by display
4295 managers such as gdm, and will go away as soon as XOrg
4296 learned native udev hotplugging for display devices.
4298 * Add new systemd-cat tool for executing arbitrary programs
4299 with STDERR/STDOUT connected to the journal. Can also act as
4300 BSD logger replacement, and does so by default.
4302 * Optionally store all locally generated coredumps in the
4303 journal along with meta data.
4305 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
4306 writing short strings to files (for usage for /sys), and for
4307 creating symlinks, character and block device nodes.
4309 * New unit file option ControlGroupPersistent= to make cgroups
4310 persistent, following the mechanisms outlined in
4311 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
4313 * Support multiple local RTCs in a sane way
4315 * No longer monopolize IO when replaying readahead data on
4316 rotating disks, since we might starve non-file-system IO to
4317 death, since fanotify() will not see accesses done by blkid,
4320 * Do not show kernel threads in systemd-cgls anymore, unless
4321 requested with new -k switch.
4323 Contributions from: Dan Horák, Kay Sievers, Lennart
4324 Poettering, Michal Schmidt
4328 * This is mostly a test release, but incorporates many
4331 * The git repository moved to:
4332 git://anongit.freedesktop.org/systemd/systemd
4333 ssh://git.freedesktop.org/git/systemd/systemd
4335 * First release with the journal
4336 http://0pointer.de/blog/projects/the-journal.html
4338 * The journal replaces both systemd-kmsg-syslogd and
4339 systemd-stdout-bridge.
4341 * New sd_pid_get_unit() API call in libsystemd-logind
4343 * Many systemadm clean-ups
4345 * Introduce remote-fs-pre.target which is ordered before all
4346 remote mounts and may be used to start services before all
4349 * Added Mageia support
4351 * Add bash completion for systemd-loginctl
4353 * Actively monitor PID file creation for daemons which exit in
4354 the parent process before having finished writing the PID
4355 file in the daemon process. Daemons which do this need to be
4356 fixed (i.e. PID file creation must have finished before the
4357 parent exits), but we now react a bit more gracefully to them.
4359 * Add colourful boot output, mimicking the well-known output
4360 of existing distributions.
4362 * New option PassCredentials= for socket units, for
4363 compatibility with a recent kernel ABI breakage.
4365 * /etc/rc.local is now hooked in via a generator binary, and
4366 thus will no longer act as synchronization point during
4369 * systemctl list-unit-files now supports --root=.
4371 * systemd-tmpfiles now understands two new commands: z, Z for
4372 relabelling files according to the SELinux database. This is
4373 useful to apply SELinux labels to specific files in /sys,
4376 * Output of SysV services is now forwarded to both the console
4377 and the journal by default, not only just the console.
4379 * New man pages for all APIs from libsystemd-login.
4381 * The build tree got reorganized and a the build system is a
4382 lot more modular allowing embedded setups to specifically
4383 select the components of systemd they are interested in.
4385 * Support for Linux systems lacking the kernel VT subsystem is
4388 * configure's --with-rootdir= got renamed to
4389 --with-rootprefix= to follow the naming used by udev and
4392 * Unless specified otherwise we will now install to /usr instead
4393 of /usr/local by default.
4395 * Processes with '@' in argv[0][0] are now excluded from the
4396 final shut-down killing spree, following the logic explained
4398 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
4400 * All processes remaining in a service cgroup when we enter
4401 the START or START_PRE states are now killed with
4402 SIGKILL. That means it is no longer possible to spawn
4403 background processes from ExecStart= lines (which was never
4404 supported anyway, and bad style).
4406 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
4407 reloading of units together.
4409 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
4410 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
4411 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
4412 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
4413 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek