1 systemd System and Service Manager
3 CHANGES WITH 243 in spe:
5 * This release enables unprivileged programs (i.e. requiring neither
6 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
7 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
8 kernel for the whole UNIX group range, i.e. all processes. This
9 change should be reasonably safe, as the kernel support for it was
10 specifically implemented to allow safe access to ICMP Echo for
11 processes lacking any privileges. If this is not desirable, it can be
12 disabled again by setting the parameter to "1 0".
14 * Previously, filters defined with SystemCallFilter= would have the
15 effect that any calling of an offending system call would terminate
16 the calling thread. This behaviour never made much sense, since
17 killing individual threads of unsuspecting processes is likely to
18 create more problems than it solves. With this release the default
19 action changed from killing the thread to killing the whole
20 process. For this to work correctly both a kernel version (>= 4.14)
21 and a libseccomp version (>= 2.4.0) supporting this new seccomp
22 action is required. If an older kernel or libseccomp is used the old
23 behaviour continues to be used. This change does not affect any
24 services that have no system call filters defined, or that use
25 SystemCallErrorNumber= (and thus see EPERM or another error instead
26 of being killed when calling an offending system call). Note that
27 systemd documentation always claimed that the whole process is
28 killed. With this change behaviour is thus adjusted to match the
31 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
32 4194304 by default, i.e. the full 22bit range the kernel allows, up
33 from the old 16bit range. This should improve security and
34 robustness, as PID collisions are made less likely (though certainly
35 still possible). There are rumours this might create compatibility
36 problems, though at this moment no practical ones are known to
37 us. Downstream distributions are hence advised to undo this change in
38 their builds if they are concerned about maximum compatibility, but
39 for everybody else we recommend leaving the value bumped. Besides
40 improving security and robustness this should also simplify things as
41 the maximum number of allowed concurrent tasks was previously bounded
42 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
43 only a single knob is left ("kernel.threads-max"). There have been
44 concerns that usability is affected by this change because larger PID
45 numbers are harder to type, but we believe the change from 5 digits
46 to 7 digits doesn't hamper usability.
48 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
49 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
50 hierarchically set default memory protection values for a particular
51 subtree of the unit hierarchy.
53 * Memory protection directives can now take a value of zero, allowing
54 explicit opting out of a default value propagated by an ancestor.
56 * systemd now defaults to the "unified" cgroup hierarchy setup during
57 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
58 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
59 change reflects the fact that cgroupsv2 support has matured
60 substantially in both systemd and in the kernel, and is clearly the
61 way forward. Downstream production distributions might want to
62 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
63 their builds as unfortunately the popular container managers have not
64 caught up with the kernel API changes.
66 * Man pages are not built by default anymore (html pages were already
67 disabled by default), to make development builds quicker. When
68 building systemd for a full installation with documentation, meson
69 should be called with -Dman=true and/or -Dhtml=true as appropriate.
70 The default was changed based on the assumption that quick one-off or
71 repeated development builds are much more common than full optimized
72 builds for installation, and people need to pass various other
73 options to when doing "proper" builds anyway, so the gain from making
74 development builds quicker is bigger than the one time disruption for
77 Two scripts are created in the *build* directory to generate and
78 preview man and html pages on demand, e.g.:
80 build/man/man systemctl
81 build/man/html systemd.index
83 * libidn2 is used by default if both libidn2 and libidn are installed.
84 Please use -Dlibidn=true if libidn is preferred.
86 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
87 big-endian machines. Before, bytes were written and read in native
88 machine order as exposed by the native libc __cpu_mask interface.
89 Now, little-endian order is always used (CPUs 0–7 are described by
90 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
91 This change fixes D-Bus calls that cross endianness boundary.
93 The presentation format used for CPUAffinity= by "systemctl show" and
94 "systemd-analyze dump" is changed to present CPU indices instead of
95 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
96 shown as CPUAffinity=03000000000000000000000000000… (on
97 little-endian) or CPUAffinity=00000000000000300000000000000… (on
98 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
99 input format. The maximum integer that will be printed in the new
100 format is 8191 (four digits), while the old format always used a very
101 long number (with the length varying by architecture), so they can be
102 unambiguously distinguished.
104 * /usr/sbin/halt.local is no longer supported. Implementation in
105 distributions was inconsistent and it seems this functionality was
108 To replace this functionality, users should:
109 - either define a new unit and make it a dependency of final.target
110 (systemctl add-wants final.target my-halt-local.service)
111 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
112 and ensure that it accepts "halt", "poweroff", "reboot", and
113 "kexec" as an argument, see the description in systemd-shutdown(8).
115 * When a [Match] section in .link or .network file is empty (contains
116 no match patterns), a warning will be emitted. Please add any "match
117 all" pattern instead, e.g. OriginalName=* or Name=* in case all
118 interfaces should really be matched.
120 * A new setting NUMAPolicy= may be used to set process memory
121 allocation policy. This setting can be specified in
122 /etc/systemd/system.conf and hence will set the default policy for
123 PID1. The default policy can be overridden on a per-service
124 basis. The related setting NUMAMask= is used to specify NUMA node
125 mask that should be associated with the selected policy.
127 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
128 generates when processes it manages are reaching their memory limits,
129 and will place their units in a special state, and optionally kill or
132 * The service manager will now expose bus properties for the IO
133 resources used by units. This information is also shown in "systemctl
134 status" now (for services that have IOAccounting=yes set). Moreover,
135 the IO accounting data is included in the resource log message
136 generated whenever a unit stops.
138 * Units may now configure an explicit time-out to wait for when killed
139 with SIGABRT, for example when a service watchdog is hit. Previously,
140 the regular TimeoutStopSec= time-out was applied in this case too —
141 now a separate time-out may be set using TimeoutAbortSec=.
143 * Services may now send a special WATCHDOG=trigger message with
144 sd_notify() to trigger an immediate "watchdog missed" event, and thus
145 trigger service termination. This is useful both for testing watchdog
146 handling, but also for defining error paths in services, that shall
147 be handled the same way as watchdog events.
149 * There are two new per-unit settings IPIngressFilterPath= and
150 IPEgressFilterPath= which allow configuration of a BPF program
151 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
152 to apply to the IP packet ingress/egress path of all processes of a
153 unit. This is useful to allow running systemd services with BPF
154 programs set up externally.
156 * systemctl gained a new "clean" verb for removing the state, cache,
157 runtime or logs directories of a service while it is terminated. The
158 new verb may also be used to remove the state maintained on disk for
159 timer units that have Persistent= configured.
161 * During the last phase of shutdown systemd will now automatically
162 increase the log level configured in the "kernel.printk" sysctl so
163 that any relevant loggable events happening during late shutdown are
164 made visible. Previously, loggable events happening so late during
165 shutdown were generally lost if the "kernel.printk" sysctl was set to
166 high thresholds, as regular logging daemons are terminated at that
167 time and thus nothing is written to disk.
169 * If processes terminated during the last phase of shutdown do not exit
170 quickly systemd will now show their names after a short time, to make
171 debugging easier. After a longer time-out they are forcibly killed,
174 * journalctl (and the other tools that display logs) will now highlight
175 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
176 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
177 are now shown in blue color, to separate them visually from regular
178 logs. References to configuration files are now turned into clickable
179 links on terminals that support that.
181 * systemd-journald will now stop logging to /var/log/journal during
182 shutdown when /var/ is on a separate mount, so that it can be
183 unmounted safely during shutdown.
185 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
187 * systemd-resolved "Cache=" configuration option in resolved.conf has
188 been extended to also accept the 'no-negative' value. Previously,
189 only a boolean option was allowed (yes/no), having yes as the
190 default. If this option is set to 'no-negative', negative answers are
191 not cached while the old cache heuristics are used positive answers.
192 The default remains unchanged.
194 * The predictable naming scheme for network devices now supports
195 generating predictable names for "netdevsim" devices.
197 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
200 Those two changes form a new net.naming-policy-scheme= entry.
201 Distributions which want to preserve naming stability may want to set
202 the -Ddefault-net-naming-scheme= configuration option.
204 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
207 * systemd-networkd's bridge FDB support now allows configuration of a
208 destination address for each entry (Destination=), as well as the
209 VXLAN VNI (VNI=), as well as an option to declare what an entry is
210 associated with (AssociatedWith=).
212 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
213 option for configuring the maximum number of DHCP lease requests. It
214 also learnt a new BlackList= option for blacklisting DHCP servers (a
215 similar setting has also been added to the IPv6 RA client), as well
216 as a SendRelease= option for configuring whether to send a DHCP
217 RELEASE message when terminating.
219 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
220 separately in the [DHCPv4] and [DHCPv6] sections.
222 * systemd-networkd's DHCP support will now optionally create an
223 implicit host route to the DNS server specified in the DHCP lease, in
224 addition to the routes listed explicitly in the lease. This should
225 ensure that in multi-homed systems DNS traffic leaves the systems on
226 the interface that acquired the DNS server information even if other
227 routes such as default routes exist. This behaviour may be turned on
228 with the new RoutesToDNS= option.
230 * systemd-networkd's VXLAN support gained a new option
231 GenericProtocolExtension= for enabling VXLAN Generic Protocol
232 Extension support, as well as IPDoNotFragment= for setting the IP
233 "Don't fragment" bit on outgoing packets. A similar option has been
234 added to the GENEVE support.
236 * In systemd-networkd's [Route] section you may now configure
237 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
238 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
239 propagation. The Type= setting now supports local, broadcast,
240 anycast, multicast, any, xresolve routes, too.
242 * systemd-networkd's [Network] section learnt a new option
243 DefaultRouteOnDevice= for automatically configuring a default route
244 onto the network device.
246 * systemd-networkd's bridging support gained two new options ProxyARP=
247 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
248 MulticastRouter= for configuring multicast routing behaviour. A new
249 option MulticastIGMPVersion= may be used to change bridge's multicast
250 Internet Group Management Protocol (IGMP) version.
252 * systemd-networkd's FooOverUDP support gained the ability to configure
253 local and peer IP addresses via Local= and Peer=. A new option
254 PeerPort= may be used to configure the peer's IP port.
256 * systemd-networkd's TUN support gained a new setting VnetHeader= for
257 tweaking Generic Segment Offload support.
259 * networkctl gained a new "delete" command for removing virtual network
260 devices, as well as a new "--stats" switch for showing device
263 * networkd.conf gained a new setting SpeedMeter= and
264 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
265 measured speed may be shown by 'networkctl status'.
267 * "networkctl status" now displays MTU and queue lengths, and more
268 detailed information about VXLAN and bridge devices.
270 * systemd-networkd's .network and .link files gained a new Property=
271 setting in the [Match] section, to match against devices with
272 specific udev properties.
274 * systemd-networkd's tunnel support gained a new option
275 AssignToLoopback= for selecting whether to use the loopback device
276 "lo" as underlying device.
278 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
279 been renamed to LinkLayerAddress=, and it now allows configuration of
282 * A new tool systemd-network-generator has been added that may generate
283 .network, .netdev and .link files from IP configuration specified on
284 the kernel command line in the format used by Dracut.
286 * The CriticalConnection= setting in .network files is now deprecated,
287 and replaced by a new KeepConfiguration= setting which allows more
288 detailed configuration of the IP configuration to keep in place.
290 * systemd-analyze gained a few new verbs:
292 - "systemd-analyze timestamp" parses and converts timestamps. This is
293 similar to the existing "systemd-analyze calendar" command which
294 does the same for recurring calendar events.
296 - "systemd-analyze timespan" parses and converts timespans (i.e.
297 durations as opposed to points in time).
299 - "systemd-analyze condition" will parse and test ConditionXYZ=
302 - "systemd-analyze exit-status" will parse and convert exit status
303 codes to their names and back.
305 - "systemd-analyze unit-files" will print a list of all unit
306 file paths and unit aliases.
308 * SuccessExitStatus=, RestartPreventExitStatus=, and
309 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
310 is equivalent to "65"). Those exit status name mappings may be
311 displayed with the sytemd-analyze exit-status verb describe above.
313 * systemd-logind now exposes a per-session SetBrightness() bus call,
314 which may be used to securely change the brightness of a kernel
315 brightness device, if it belongs to the session's seat. By using this
316 call unprivileged clients can make changes to "backlight" and "leds"
317 devices securely with strict requirements on session membership.
318 Desktop environments may use this to generically make brightness
319 changes to such devices without shipping private SUID binaries or
320 udev rules for that purpose.
322 * "udevadm info" gained a --wait-for-initialization switch to wait for
323 a device to be initialized.
325 * systemd-hibernate-resume-generator will now look for resumeflags= on
326 the kernel command line, which is similar to rootflags= and may be
327 used to configure device timeout for the hibernation device.
329 * sd-event learnt a new API call sd_event_source_disable_unref() for
330 disabling and unref'ing an event source in a single function. A
331 related call sd_event_source_disable_unrefp() has been added for use
332 with gcc's cleanup extension.
334 * The sd-id128.h public API gained a new definition
335 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
338 * "busctl introspect" gained a new switch --xml-interface for dumping
339 XML introspection data unmodified.
341 * PID 1 may now show the unit name instead of the unit description
342 string in its status output during boot. This may be configured in
343 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
344 kernel command line option systemd.status_unit_format=.
346 * PID 1 now understands a new option KExecWatchdogSec= in
347 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
348 Previously watchdog functionality was only available for regular
349 reboots. The new setting defaults to off, because we don't know in
350 the general case if the watchdog will be reset after kexec (some
351 drivers do reset it, but not all), and the new userspace might not be
352 configured to handle the watchdog.
354 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
355 RebootWatchdogSec= to more clearly communicate what it is about. The
356 old name is still accepted for compatibility.
358 * The systemd.debug_shell kernel command line option now optionally
359 takes a tty name to spawn the debug shell on, which allows a
360 different tty to be selected than the built-in default.
362 * Service units gained a new ExecCondition= setting which will run
363 before ExecStartPre= and either continue execution of the unit (for
364 clean exit codes), stop execution without marking the unit failed
365 (for exit codes 1 through 254), or stop execution and fail the unit
366 (for exit code 255 or abnormal termination).
368 * A new service systemd-pstore.service has been added that pulls data
369 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
372 * timedatectl gained new verbs for configuring per-interface NTP
373 service configuration for systemd-timesyncd.
375 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
376 2019. (You can set non-UTF-8 locales though, if you know their name.)
378 * If variable assignments in sysctl.d/ files are prefixed with "-" any
379 failures to apply them are now ignored.
381 * systemd-random-seed.service now optionally credits entropy when
382 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
383 true for the service to enable this behaviour, but please consult the
384 documentation first, since this comes with a couple of caveats.
386 * systemd-random-seed.service is now a synchronization point for full
387 initialization of the kernel's entropy pool. Services that require
388 /dev/urandom to be correctly initialized should be ordered after this
391 * The systemd-boot boot loader has been updated to optionally maintain
392 a random seed file in the EFI System Partition (ESP). During the boot
393 phase, this random seed is read and updated with a new seed
394 cryptographically derived from it. Another derived seed is passed to
395 the OS. The latter seed is then credited to the kernel's entropy pool
396 very early during userspace initialization (from PID 1). This allows
397 systems to boot up with a fully initialized kernel entropy pool from
398 earliest boot on, and thus entirely removes all entropy pool
399 initialization delays from systems using systemd-boot. Special care
400 is taken to ensure different seeds are derived on system images
401 replicated to multiple systems. "bootctl status" will show whether
402 a seed was received from the boot loader.
404 * bootctl gained two new verbs:
406 - "bootctl random-seed" will generate the file in ESP and an EFI
407 variable to allow a random seed to be passed to the OS as described
410 - "bootctl is-installed" checks whether systemd-boot is currently
413 * bootctl will warn if it detects that boot entries are misconfigured
414 (for example if the kernel image was removed without purging the
417 * A new document has been added describing systemd's use and support
418 for the kernel's entropy pool subsystem:
420 https://systemd.io/RANDOM_SEEDS
422 * When the system is hibernated the swap device to write the
423 hibernation image to is now automatically picked from all available
424 swap devices, preferring the swap device with the highest configured
425 priority over all others, and picking the device with the most free
426 space if there are multiple devices with the highest priority.
428 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
429 option that permits selecting the timout how long to wait for a
430 device with an encryption key before asking for the password.
432 * IOWeight= has learnt to properly set the IO weight when using the
433 BFQ scheduler officially found in kernels 5.0+.
435 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
436 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
437 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
438 Chiu, Chris Down, Christian Kellner, Clinton Roy, Connor Reeder, Daniel
439 Black, Daniele Medri, Dan Streetman, Dave Reisner, Dave Ross, David
440 Art, David Tardon, Debarshi Ray, Dominick Grift, Donald Buczek, Douglas
441 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
442 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
443 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
444 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
445 Unterwurzacher, Jan Klötzke, Jan Pokorný, Jan Synacek, Jeka Pats,
446 Jérémy Rosen, Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann
447 B. Guðmundsson, Johannes Christ, Johannes Schmitz, Jonathan Rouleau,
448 Jorge Niedbalski, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
449 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
450 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
451 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
452 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
453 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
454 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
455 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Roberto
456 Santalla, Ronan Pigott, root, RussianNeuroMancer, Sebastian Jennen,
457 shinygold, Shreyas Behera, Simon Schricker, Susant Sahani, Thadeu Lima
458 de Souza Cascardo, Theo Ouzhinski, Thiebaud Weksteen, Thomas Haller,
459 Thomas Weißschuh, Tomas Mraz, Tommi Rantala, Topi Miettinen, ven,
460 Wieland Hoffmann, William A. Kennington III, William Wold, Xi Ruoyao,
461 Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
464 – Somewhere, SOME-TI-ME
468 * In .link files, MACAddressPolicy=persistent (the default) is changed
469 to cover more devices. For devices like bridges, tun, tap, bond, and
470 similar interfaces that do not have other identifying information,
471 the interface name is used as the basis for persistent seed for MAC
472 and IPv4LL addresses. The way that devices that were handled
473 previously is not changed, and this change is about covering more
474 devices then previously by the "persistent" policy.
476 MACAddressPolicy=random may be used to force randomized MACs and
477 IPv4LL addresses for a device if desired.
479 Hint: the log output from udev (at debug level) was enhanced to
480 clarify what policy is followed and which attributes are used.
481 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
482 may be used to view this.
484 Hint: if a bridge interface is created without any slaves, and gains
485 a slave later, then now the bridge does not inherit slave's MAC.
486 To inherit slave's MAC, for example, create the following file:
488 # /etc/systemd/network/98-bridge-inherit-mac.link
493 MACAddressPolicy=none
496 * The .device units generated by systemd-fstab-generator and other
497 generators do not automatically pull in the corresponding .mount unit
498 as a Wants= dependency. This means that simply plugging in the device
499 will not cause the mount unit to be started automatically. But please
500 note that the mount unit may be started for other reasons, in
501 particular if it is part of local-fs.target, and any unit which
502 (transitively) depends on local-fs.target is started.
504 * networkctl list/status/lldp now accept globbing wildcards for network
505 interface names to match against all existing interfaces.
507 * The $PIDFILE environment variable is set to point the absolute path
508 configured with PIDFile= for processes of that service.
510 * The fallback DNS server list was augmented with Cloudflare public DNS
511 servers. Use `-Ddns-servers=` to set a different fallback.
513 * A new special target usb-gadget.target will be started automatically
514 when a USB Device Controller is detected (which means that the system
515 is a USB peripheral).
517 * A new unit setting CPUQuotaPeriodSec= assigns the time period
518 relatively to which the CPU time quota specified by CPUQuota= is
521 * A new unit setting ProtectHostname= may be used to prevent services
522 from modifying hostname information (even if they otherwise would
523 have privileges to do so).
525 * A new unit setting NetworkNamespacePath= may be used to specify a
526 namespace for service or socket units through a path referring to a
527 Linux network namespace pseudo-file.
529 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
530 have an effect on .socket units: when used the listening socket is
531 created within the configured network namespace instead of the host
534 * ExecStart= command lines in unit files may now be prefixed with ':'
535 in which case environment variable substitution is
536 disabled. (Supported for the other ExecXYZ= settings, too.)
538 * .timer units gained two new boolean settings OnClockChange= and
539 OnTimezoneChange= which may be used to also trigger a unit when the
540 system clock is changed or the local timezone is
541 modified. systemd-run has been updated to make these options easily
542 accessible from the command line for transient timers.
544 * Two new conditions for units have been added: ConditionMemory= may be
545 used to conditionalize a unit based on installed system
546 RAM. ConditionCPUs= may be used to conditionalize a unit based on
549 * The @default system call filter group understood by SystemCallFilter=
550 has been updated to include the new rseq() system call introduced in
553 * A new time-set.target has been added that indicates that the system
554 time has been set from a local source (possibly imprecise). The
555 existing time-sync.target is stronger and indicates that the time has
556 been synchronized with a precise external source. Services where
557 approximate time is sufficient should use the new target.
559 * "systemctl start" (and related commands) learnt a new
560 --show-transaction option. If specified brief information about all
561 jobs queued because of the requested operation is shown.
563 * systemd-networkd recognizes a new operation state 'enslaved', used
564 (instead of 'degraded' or 'carrier') for interfaces which form a
565 bridge, bond, or similar, and an new 'degraded-carrier' operational
566 state used for the bond or bridge master interface when one of the
567 enslaved devices is not operational.
569 * .network files learnt the new IgnoreCarrierLoss= option for leaving
570 networks configured even if the carrier is lost.
572 * The RequiredForOnline= setting in .network files may now specify a
573 minimum operational state required for the interface to be considered
574 "online" by systemd-networkd-wait-online. Related to this
575 systemd-networkd-wait-online gained a new option --operational-state=
576 to configure the same, and its --interface= option was updated to
577 optionally also take an operational state specific for an interface.
579 * systemd-networkd-wait-online gained a new setting --any for waiting
580 for only one of the requested interfaces instead of all of them.
582 * systemd-networkd now implements L2TP tunnels.
584 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
585 may be used to cause autonomous and onlink prefixes received in IPv6
586 Router Advertisements to be ignored.
588 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
589 file settings may be used to tweak bridge behaviour.
591 * The new TripleSampling= option in .network files may be used to
592 configure CAN triple sampling.
594 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
595 used to point to private or preshared key for a WireGuard interface.
597 * /etc/crypttab now supports the same-cpu-crypt and
598 submit-from-crypt-cpus options to tweak encryption work scheduling
601 * systemd-tmpfiles will now take a BSD file lock before operating on a
602 contents of directory. This may be used to temporarily exclude
603 directories from aging by taking the same lock (useful for example
604 when extracting a tarball into /tmp or /var/tmp as a privileged user,
605 which might create files with really old timestamps, which
606 nevertheless should not be deleted). For further details, see:
608 https://systemd.io/TEMPORARY_DIRECTORIES
610 * systemd-tmpfiles' h line type gained support for the
611 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
612 controlling project quota inheritance.
614 * sd-boot and bootctl now implement support for an Extended Boot Loader
615 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
616 addition to the ESP partition mounted to /efi or /boot/efi.
617 Configuration file fragments, kernels, initrds and other EFI images
618 to boot will be loaded from both the ESP and XBOOTLDR partitions.
619 The XBOOTLDR partition was previously described by the Boot Loader
620 Specification, but implementation was missing in sd-boot. Support for
621 this concept allows using the sd-boot boot loader in more
622 conservative scenarios where the boot loader itself is placed in the
623 ESP but the kernels to boot (and their metadata) in a separate
626 * A system may now be booted with systemd.volatile=overlay on the
627 kernel command line, which causes the root file system to be set up
628 an overlayfs mount combining the root-only root directory with a
629 writable tmpfs. In this setup, the underlying root device is not
630 modified, and any changes are lost at reboot.
632 * Similar, systemd-nspawn can now boot containers with a volatile
633 overlayfs root with the new --volatile=overlay switch.
635 * systemd-nspawn can now consume OCI runtime bundles using a new
636 --oci-bundle= option. This implementation is fully usable, with most
637 features in the specification implemented, but since this a lot of
638 new code and functionality, this feature should most likely not
639 be used in production yet.
641 * systemd-nspawn now supports various options described by the OCI
642 runtime specification on the command-line and in .nspawn files:
643 --inaccessible=/Inaccessible= may be used to mask parts of the file
644 system tree, --console=/--pipe may be used to configure how standard
645 input, output, and error are set up.
647 * busctl learned the `emit` verb to generate D-Bus signals.
649 * systemd-analyze cat-config may be used to gather and display
650 configuration spread over multiple files, for example system and user
651 presets, tmpfiles.d, sysusers.d, udev rules, etc.
653 * systemd-analyze calendar now takes an optional new parameter
654 --iterations= which may be used to show a maximum number of iterations
655 the specified expression will elapse next.
657 * The sd-bus C API gained support for naming method parameters in the
660 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
661 the reboot() system call expects.
663 * journalctl learnt a new --cursor-file= option that points to a file
664 from which a cursor should be loaded in the beginning and to which
665 the updated cursor should be stored at the end.
667 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
668 detected by systemd-detect-virt (and may also be used in
669 ConditionVirtualization=).
671 * The behaviour of systemd-logind may now be modified with environment
672 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
673 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
674 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
675 skip the relevant operation completely (when set to false), or to
676 create a flag file in /run/systemd (when set to true), instead of
677 actually commencing the real operation when requested. The presence
678 of /run/systemd/reboot-to-firmware-setup,
679 /run/systemd/reboot-to-boot-loader-menu, and
680 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
681 boot loader implementations to replace some steps logind performs
682 during reboot with their own operations.
684 * systemctl can be used to request a reboot into the boot loader menu
685 or a specific boot loader entry with the new --boot-load-menu= and
686 --boot-loader-entry= options to a reboot command. (This requires a
687 boot loader that supports this, for example sd-boot.)
689 * kernel-install will no longer unconditionally create the output
690 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
691 snippets, but will do only if the machine-specific parent directory
692 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
693 to create this parent directory during sd-boot installation.
695 This makes it easier to use kernel-install with plugins which support
696 a different layout of the bootloader partitions (for example grub2).
698 * During package installation (with `ninja install`), we would create
699 symlinks for getty@tty1.service, systemd-networkd.service,
700 systemd-networkd.socket, systemd-resolved.service,
701 remote-cryptsetup.target, remote-fs.target,
702 systemd-networkd-wait-online.service, and systemd-timesyncd.service
703 in /etc, as if `systemctl enable` was called for those units, to make
704 the system usable immediately after installation. Now this is not
705 done anymore, and instead calling `systemctl preset-all` is
706 recommended after the first installation of systemd.
708 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
709 is built on seccomp. When turned on creation of SUID/SGID files is
712 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
713 implied if DynamicUser= is turned on for a service. This hardens
714 these services, so that they neither can benefit from nor create
715 SUID/SGID executables. This is a minor compatibility breakage, given
716 that when DynamicUser= was first introduced SUID/SGID behaviour was
717 unaffected. However, the security benefit of these two options is
718 substantial, and the setting is still relatively new, hence we opted
719 to make it mandatory for services with dynamic users.
721 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
722 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
723 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
724 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
725 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
726 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
727 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
728 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
729 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
730 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
731 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
732 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
733 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
734 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
735 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
736 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
737 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
738 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
744 * The default locale can now be configured at compile time. Otherwise,
745 a suitable default will be selected automatically (one of C.UTF-8,
748 * The version string shown by systemd and other tools now includes the
749 git commit hash when built from git. An override may be specified
750 during compilation, which is intended to be used by distributions to
751 include the package release information.
753 * systemd-cat can now filter standard input and standard error streams
754 for different syslog priorities using the new --stderr-priority=
757 * systemd-journald and systemd-journal-remote reject entries which
758 contain too many fields (CVE-2018-16865) and set limits on the
759 process' command line length (CVE-2018-16864).
761 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
764 * A new network device NamePolicy "keep" is implemented for link files,
765 and used by default in 99-default.link (the fallback configuration
766 provided by systemd). With this policy, if the network device name
767 was already set by userspace, the device will not be renamed again.
768 This matches the naming scheme that was implemented before
769 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
770 is also enabled by default, even if not specified. Effectively, this
771 means that if naming-scheme >= 240 is specified, network devices will
772 be renamed according to the configuration, even if they have been
773 renamed already, if "keep" is not specified as the naming policy in
774 the .link file. The 99-default.link file provided by systemd includes
775 "keep" for backwards compatibility, but it is recommended for user
776 installed .link files to *not* include it.
778 The "kernel" policy, which keeps kernel names declared to be
779 "persistent", now works again as documented.
781 * kernel-install script now optionally takes the paths to one or more
782 initrd files, and passes them to all plugins.
784 * The mincore() system call has been dropped from the @system-service
785 system call filter group, as it is pretty exotic and may potentially
786 used for side-channel attacks.
788 * -fPIE is dropped from compiler and linker options. Please specify
789 -Db_pie=true option to meson to build position-independent
790 executables. Note that the meson option is supported since meson-0.49.
792 * The fs.protected_regular and fs.protected_fifos sysctls, which were
793 added in Linux 4.19 to make some data spoofing attacks harder, are
794 now enabled by default. While this will hopefully improve the
795 security of most installations, it is technically a backwards
796 incompatible change; to disable these sysctls again, place the
797 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
799 fs.protected_regular = 0
800 fs.protected_fifos = 0
802 Note that the similar hardlink and symlink protection has been
803 enabled since v199, and may be disabled likewise.
805 * The files read from the EnvironmentFile= setting in unit files now
806 parse backslashes inside quotes literally, matching the behaviour of
809 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
810 now automatically become NOPs when run in a chroot() environment.
812 * The tmpfiles.d/ "C" line type will now copy directory trees not only
813 when the destination is so far missing, but also if it already exists
814 as a directory and is empty. This is useful to cater for systems
815 where directory trees are put together from multiple separate mount
816 points but otherwise empty.
818 * A new function sd_bus_close_unref() (and the associated
819 sd_bus_close_unrefp()) has been added to libsystemd, that combines
820 sd_bus_close() and sd_bus_unref() in one.
822 * udevadm control learnt a new option for --ping for testing whether a
823 systemd-udevd instance is running and reacting.
825 * udevadm trigger learnt a new option for --wait-daemon for waiting
826 systemd-udevd daemon to be initialized.
828 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
829 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
830 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
831 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
832 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
833 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
834 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
835 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
836 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
837 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
838 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
839 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
840 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
841 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
842 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
843 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
844 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
850 * NoNewPrivileges=yes has been set for all long-running services
851 implemented by systemd. Previously, this was problematic due to
852 SELinux (as this would also prohibit the transition from PID1's label
853 to the service's label). This restriction has since been lifted, but
854 an SELinux policy update is required.
855 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
857 * DynamicUser=yes is dropped from systemd-networkd.service,
858 systemd-resolved.service and systemd-timesyncd.service, which was
859 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
860 and since v236 for systemd-timesyncd.service. The users and groups
861 systemd-network, systemd-resolve and systemd-timesync are created
862 by systemd-sysusers again. Distributors or system administrators
863 may need to create these users and groups if they not exist (or need
864 to re-enable DynamicUser= for those units) while upgrading systemd.
865 Also, the clock file for systemd-timesyncd may need to move from
866 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
868 * When unit files are loaded from disk, previously systemd would
869 sometimes (depending on the unit loading order) load units from the
870 target path of symlinks in .wants/ or .requires/ directories of other
871 units. This meant that unit could be loaded from different paths
872 depending on whether the unit was requested explicitly or as a
873 dependency of another unit, not honouring the priority of directories
874 in search path. It also meant that it was possible to successfully
875 load and start units which are not found in the unit search path, as
876 long as they were requested as a dependency and linked to from
877 .wants/ or .requires/. The target paths of those symlinks are not
878 used for loading units anymore and the unit file must be found in
881 * A new service type has been added: Type=exec. It's very similar to
882 Type=simple but ensures the service manager will wait for both fork()
883 and execve() of the main service binary to complete before proceeding
884 with follow-up units. This is primarily useful so that the manager
885 propagates any errors in the preparation phase of service execution
886 back to the job that requested the unit to be started. For example,
887 consider a service that has ExecStart= set to a file system binary
888 that doesn't exist. With Type=simple starting the unit would be
889 considered instantly successful, as only fork() has to complete
890 successfully and the manager does not wait for execve(), and hence
891 its failure is seen "too late". With the new Type=exec service type
892 starting the unit will fail, as the manager will wait for the
893 execve() and notice its failure, which is then propagated back to the
896 NOTE: with the next release 241 of systemd we intend to change the
897 systemd-run tool to default to Type=exec for transient services
898 started by it. This should be mostly safe, but in specific corner
899 cases might result in problems, as the systemd-run tool will then
900 block on NSS calls (such as user name look-ups due to User=) done
901 between the fork() and execve(), which under specific circumstances
902 might cause problems. It is recommended to specify "-p Type=simple"
903 explicitly in the few cases where this applies. For regular,
904 non-transient services (i.e. those defined with unit files on disk)
905 we will continue to default to Type=simple.
907 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
908 userspace processes is set to 1024 (soft) and 4096
909 (hard). Previously, systemd passed this on unmodified to all
910 processes it forked off. With this systemd release the hard limit
911 systemd passes on is increased to 512K, overriding the kernel's
912 defaults and substantially increasing the number of simultaneous file
913 descriptors unprivileged userspace processes can allocate. Note that
914 the soft limit remains at 1024 for compatibility reasons: the
915 traditional UNIX select() call cannot deal with file descriptors >=
916 1024 and increasing the soft limit globally might thus result in
917 programs unexpectedly allocating a high file descriptor and thus
918 failing abnormally when attempting to use it with select() (of
919 course, programs shouldn't use select() anymore, and prefer
920 poll()/epoll, but the call unfortunately remains undeservedly popular
921 at this time). This change reflects the fact that file descriptor
922 handling in the Linux kernel has been optimized in more recent
923 kernels and allocating large numbers of them should be much cheaper
924 both in memory and in performance than it used to be. Programs that
925 want to take benefit of the increased limit have to "opt-in" into
926 high file descriptors explicitly by raising their soft limit. Of
927 course, when they do that they must acknowledge that they cannot use
928 select() anymore (and neither can any shared library they use — or
929 any shared library used by any shared library they use and so on).
930 Which default hard limit is most appropriate is of course hard to
931 decide. However, given reports that ~300K file descriptors are used
932 in real-life applications we believe 512K is sufficiently high as new
933 default for now. Note that there are also reports that using very
934 high hard limits (e.g. 1G) is problematic: some software allocates
935 large arrays with one element for each potential file descriptor
936 (Java, …) — a high hard limit thus triggers excessively large memory
937 allocations in these applications. Hopefully, the new default of 512K
938 is a good middle ground: higher than what real-life applications
939 currently need, and low enough for avoid triggering excessively large
940 allocations in problematic software. (And yes, somebody should fix
943 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
944 to the highest possible values, as separate accounting of file
945 descriptors is no longer necessary, as memcg tracks them correctly as
946 part of the memory accounting anyway. Thus, from the four limits on
947 file descriptors currently enforced (fs.file-max, fs.nr_open,
948 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
949 and keep only the latter two. A set of build-time options
950 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
951 has been added to revert this change in behaviour, which might be
952 an option for systems that turn off memcg in the kernel.
954 * When no /etc/locale.conf file exists (and hence no locale settings
955 are in place), systemd will now use the "C.UTF-8" locale by default,
956 and set LANG= to it. This locale is supported by various
957 distributions including Fedora, with clear indications that upstream
958 glibc is going to make it available too. This locale enables UTF-8
959 mode by default, which appears appropriate for 2018.
961 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
962 default. This effectively switches the RFC3704 Reverse Path filtering
963 from Strict mode to Loose mode. This is more appropriate for hosts
964 that have multiple links with routes to the same networks (e.g.
965 a client with a Wi-Fi and Ethernet both connected to the internet).
967 Consult the kernel documentation for details on this sysctl:
968 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
970 * CPUAccounting=yes no longer enables the CPU controller when using
971 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
972 statistics are now provided independently from the CPU controller.
974 * Support for disabling a particular cgroup controller within a sub-tree
975 has been added through the DisableControllers= directive.
977 * cgroup_no_v1=all on the kernel command line now also implies
978 using the unified cgroup hierarchy, unless one explicitly passes
979 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
981 * The new "MemoryMin=" unit file property may now be used to set the
982 memory usage protection limit of processes invoked by the unit. This
983 controls the cgroup v2 memory.min attribute. Similarly, the new
984 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
985 cgroup v2 io.latency cgroup property for configuring per-service I/O
988 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
989 to the cgroup v1 "devices" cgroup controller.
991 * systemd-escape now is able to combine --unescape with --template. It
992 also learnt a new option --instance for extracting and unescaping the
993 instance part of a unit name.
995 * sd-bus now provides the sd_bus_message_readv() which is similar to
996 sd_bus_message_read() but takes a va_list object. The pair
997 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
998 has been added for configuring the default method call timeout to
999 use. sd_bus_error_move() may be used to efficiently move the contents
1000 from one sd_bus_error structure to another, invalidating the
1001 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
1002 be used to control whether a bus connection object is automatically
1003 flushed when an sd-event loop is exited.
1005 * When processing classic BSD syslog log messages, journald will now
1006 save the original time-stamp string supplied in the new
1007 SYSLOG_TIMESTAMP= journal field. This permits consumers to
1008 reconstruct the original BSD syslog message more correctly.
1010 * StandardOutput=/StandardError= in service files gained support for
1011 new "append:…" parameters, for connecting STDOUT/STDERR of a service
1012 to a file, and appending to it.
1014 * The signal to use as last step of killing of unit processes is now
1015 configurable. Previously it was hard-coded to SIGKILL, which may now
1016 be overridden with the new KillSignal= setting. Note that this is the
1017 signal used when regular termination (i.e. SIGTERM) does not suffice.
1018 Similarly, the signal used when aborting a program in case of a
1019 watchdog timeout may now be configured too (WatchdogSignal=).
1021 * The XDG_SESSION_DESKTOP environment variable may now be configured in
1022 the pam_systemd argument line, using the new desktop= switch. This is
1023 useful to initialize it properly from a display manager without
1024 having to touch C code.
1026 * Most configuration options that previously accepted percentage values
1027 now also accept permille values with the '‰' suffix (instead of '%').
1029 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
1032 * systemd-resolved's configuration file resolved.conf gained a new
1033 option ReadEtcHosts= which may be used to turn off processing and
1034 honoring /etc/hosts entries.
1036 * The "--wait" switch may now be passed to "systemctl
1037 is-system-running", in which case the tool will synchronously wait
1038 until the system finished start-up.
1040 * hostnamed gained a new bus call to determine the DMI product UUID.
1042 * On x86-64 systemd will now prefer using the RDRAND processor
1043 instruction over /dev/urandom whenever it requires randomness that
1044 neither has to be crypto-grade nor should be reproducible. This
1045 should substantially reduce the amount of entropy systemd requests
1046 from the kernel during initialization on such systems, though not
1047 reduce it to zero. (Why not zero? systemd still needs to allocate
1048 UUIDs and such uniquely, which require high-quality randomness.)
1050 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
1051 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
1052 for forcing the "Other Information" bit in IPv6 RA messages. The
1053 bonding logic gained four new options AdActorSystemPriority=,
1054 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
1055 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
1056 shuffling of flows. The tunnel logic gained a new
1057 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
1058 Deployment. The policy rule logic gained four new options IPProtocol=,
1059 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
1060 support for the MulticastToUnicast= option. networkd also gained
1061 support for configuring static IPv4 ARP or IPv6 neighbor entries.
1063 * .preset files (as read by 'systemctl preset') may now be used to
1064 instantiate services.
1066 * /etc/crypttab now understands the sector-size= option to configure
1067 the sector size for an encrypted partition.
1069 * Key material for encrypted disks may now be placed on a formatted
1070 medium, and referenced from /etc/crypttab by the UUID of the file
1071 system, followed by "=" suffixed by the path to the key file.
1073 * The "collect" udev component has been removed without replacement, as
1074 it is neither used nor maintained.
1076 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
1077 LogsDirectory=, ConfigurationDirectory= settings are used in a
1078 service the executed processes will now receive a set of environment
1079 variables containing the full paths of these directories.
1080 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
1081 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
1082 are used. Note that these options may be used multiple times per
1083 service in which case the resulting paths will be concatenated and
1084 separated by colons.
1086 * Predictable interface naming has been extended to cover InfiniBand
1087 NICs. They will be exposed with an "ib" prefix.
1089 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
1090 which case the respective line failing is ignored.
1092 * .link files may now be used to configure the equivalent to the
1093 "ethtool advertise" commands.
1095 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
1096 alternative to libudev.h. Previously, the latter was just an internal
1097 wrapper around the former, but now these two APIs are exposed
1100 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
1101 which calculates an app-specific boot ID similar to how
1102 sd_id128_get_machine_app_specific() generates an app-specific machine
1105 * A new tool systemd-id128 has been added that can be used to determine
1106 and generate various 128bit IDs.
1108 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
1111 * systemd-hibernate-resume-generator will now honor the "noresume"
1112 kernel command line option, in which case it will bypass resuming
1113 from any hibernated image.
1115 * The systemd-sleep.conf configuration file gained new options
1116 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
1117 AllowHybridSleep= for prohibiting specific sleep modes even if the
1118 kernel exports them.
1120 * portablectl is now officially supported and has thus moved to
1123 * bootctl learnt the two new commands "set-default" and "set-oneshot"
1124 for setting the default boot loader item to boot to (either
1125 persistently or only for the next boot). This is currently only
1126 compatible with sd-boot, but may be implemented on other boot loaders
1127 too, that follow the boot loader interface. The updated interface is
1128 now documented here:
1130 https://systemd.io/BOOT_LOADER_INTERFACE
1132 * A new kernel command line option systemd.early_core_pattern= is now
1133 understood which may be used to influence the core_pattern PID 1
1134 installs during early boot.
1136 * busctl learnt two new options -j and --json= for outputting method
1137 call replies, properties and monitoring output in JSON.
1139 * journalctl's JSON output now supports simple ANSI coloring as well as
1140 a new "json-seq" mode for generating RFC7464 output.
1142 * Unit files now support the %g/%G specifiers that resolve to the UNIX
1143 group/GID of the service manager runs as, similar to the existing
1144 %u/%U specifiers that resolve to the UNIX user/UID.
1146 * systemd-logind learnt a new global configuration option
1147 UserStopDelaySec= that may be set in logind.conf. It specifies how
1148 long the systemd --user instance shall remain started after a user
1149 logs out. This is useful to speed up repetitive re-connections of the
1150 same user, as it means the user's service manager doesn't have to be
1151 stopped/restarted on each iteration, but can be reused between
1152 subsequent options. This setting defaults to 10s. systemd-logind also
1153 exports two new properties on its Manager D-Bus objects indicating
1154 whether the system's lid is currently closed, and whether the system
1157 * systemd gained support for a generic boot counting logic, which
1158 generically permits automatic reverting to older boot loader entries
1159 if newer updated ones don't work. The boot loader side is implemented
1160 in sd-boot, but is kept open for other boot loaders too. For details
1163 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
1165 * The SuccessAction=/FailureAction= unit file settings now learnt two
1166 new parameters: "exit" and "exit-force", which result in immediate
1167 exiting of the service manager, and are only useful in systemd --user
1168 and container environments.
1170 * Unit files gained support for a pair of options
1171 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
1172 exit status to use as service manager exit status when
1173 SuccessAction=/FailureAction= is set to exit or exit-force.
1175 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
1176 options may now be used to configure the log rate limiting applied by
1177 journald per-service.
1179 * systemd-analyze gained a new verb "timespan" for parsing and
1180 normalizing time span values (i.e. strings like "5min 7s 8us").
1182 * systemd-analyze also gained a new verb "security" for analyzing the
1183 security and sand-boxing settings of services in order to determine an
1184 "exposure level" for them, indicating whether a service would benefit
1185 from more sand-boxing options turned on for them.
1187 * "systemd-analyze syscall-filter" will now also show system calls
1188 supported by the local kernel but not included in any of the defined
1191 * .nspawn files now understand the Ephemeral= setting, matching the
1192 --ephemeral command line switch.
1194 * sd-event gained the new APIs sd_event_source_get_floating() and
1195 sd_event_source_set_floating() for controlling whether a specific
1196 event source is "floating", i.e. destroyed along with the even loop
1199 * Unit objects on D-Bus gained a new "Refs" property that lists all
1200 clients that currently have a reference on the unit (to ensure it is
1203 * The JoinControllers= option in system.conf is no longer supported, as
1204 it didn't work correctly, is hard to support properly, is legacy (as
1205 the concept only exists on cgroup v1) and apparently wasn't used.
1207 * Journal messages that are generated whenever a unit enters the failed
1208 state are now tagged with a unique MESSAGE_ID. Similarly, messages
1209 generated whenever a service process exits are now made recognizable,
1210 too. A tagged message is also emitted whenever a unit enters the
1211 "dead" state on success.
1213 * systemd-run gained a new switch --working-directory= for configuring
1214 the working directory of the service to start. A shortcut -d is
1215 equivalent, setting the working directory of the service to the
1216 current working directory of the invoking program. The new --shell
1217 (or just -S) option has been added for invoking the $SHELL of the
1218 caller as a service, and implies --pty --same-dir --wait --collect
1219 --service-type=exec. Or in other words, "systemd-run -S" is now the
1220 quickest way to quickly get an interactive in a fully clean and
1221 well-defined system service context.
1223 * machinectl gained a new verb "import-fs" for importing an OS tree
1224 from a directory. Moreover, when a directory or tarball is imported
1225 and single top-level directory found with the OS itself below the OS
1226 tree is automatically mangled and moved one level up.
1228 * systemd-importd will no longer set up an implicit btrfs loop-back
1229 file system on /var/lib/machines. If one is already set up, it will
1230 continue to be used.
1232 * A new generator "systemd-run-generator" has been added. It will
1233 synthesize a unit from one or more program command lines included in
1234 the kernel command line. This is very useful in container managers
1237 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
1239 This will run "systemd-nspawn" on an image, invoke the specified
1240 command line and immediately shut down the container again, returning
1241 the command line's exit code.
1243 * The block device locking logic is now documented:
1245 https://systemd.io/BLOCK_DEVICE_LOCKING
1247 * loginctl and machinectl now optionally output the various tables in
1248 JSON using the --output= switch. It is our intention to add similar
1249 support to systemctl and all other commands.
1251 * udevadm's query and trigger verb now optionally take a .device unit
1254 * systemd-udevd's network naming logic now understands a new
1255 net.naming-scheme= kernel command line switch, which may be used to
1256 pick a specific version of the naming scheme. This helps stabilizing
1257 interface names even as systemd/udev are updated and the naming logic
1260 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
1261 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
1262 initialize one to all 0xFF.
1264 * After loading the SELinux policy systemd will now recursively relabel
1265 all files and directories listed in
1266 /run/systemd/relabel-extra.d/*.relabel (which should be simple
1267 newline separated lists of paths) in addition to the ones it already
1268 implicitly relabels in /run, /dev and /sys. After the relabelling is
1269 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
1270 removed. This is useful to permit initrds (i.e. code running before
1271 the SELinux policy is in effect) to generate files in the host
1272 filesystem safely and ensure that the correct label is applied during
1273 the transition to the host OS.
1275 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
1276 mknod() handling in user namespaces. Previously mknod() would always
1277 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
1278 but device nodes generated that way cannot be opened, and attempts to
1279 open them result in EPERM. This breaks the "graceful fallback" logic
1280 in systemd's PrivateDevices= sand-boxing option. This option is
1281 implemented defensively, so that when systemd detects it runs in a
1282 restricted environment (such as a user namespace, or an environment
1283 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
1284 where device nodes cannot be created the effect of PrivateDevices= is
1285 bypassed (following the logic that 2nd-level sand-boxing is not
1286 essential if the system systemd runs in is itself already sand-boxed
1287 as a whole). This logic breaks with 4.18 in container managers where
1288 user namespacing is used: suddenly PrivateDevices= succeeds setting
1289 up a private /dev/ file system containing devices nodes — but when
1290 these are opened they don't work.
1292 At this point it is recommended that container managers utilizing
1293 user namespaces that intend to run systemd in the payload explicitly
1294 block mknod() with seccomp or similar, so that the graceful fallback
1297 We are very sorry for the breakage and the requirement to change
1298 container configurations for newer kernels. It's purely caused by an
1299 incompatible kernel change. The relevant kernel developers have been
1300 notified about this userspace breakage quickly, but they chose to
1303 * PermissionsStartOnly= setting is deprecated (but is still supported
1304 for backwards compatibility). The same functionality is provided by
1305 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
1308 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
1309 pam_systemd anymore.
1311 * The naming scheme for network devices was changed to always rename
1312 devices, even if they were already renamed by userspace. The "kernel"
1313 policy was changed to only apply as a fallback, if no other naming
1316 * The requirements to build systemd is bumped to meson-0.46 and
1319 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
1320 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
1321 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
1322 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
1323 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
1324 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
1325 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
1326 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
1327 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
1328 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
1329 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
1330 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
1331 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
1332 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
1333 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
1334 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
1335 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
1336 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
1337 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
1338 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
1339 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
1340 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
1341 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
1342 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
1343 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
1344 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
1345 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
1346 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
1347 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
1348 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
1349 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
1350 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
1351 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
1352 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
1353 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
1354 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
1355 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
1356 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
1357 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
1358 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
1359 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
1360 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
1361 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
1362 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
1363 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
1365 — Warsaw, 2018-12-21
1369 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
1370 builtin will name network interfaces differently than in previous
1371 versions for virtual network interfaces created with SR-IOV and NPAR
1372 and for devices where the PCI network controller device does not have
1373 a slot number associated.
1375 SR-IOV virtual devices are now named based on the name of the parent
1376 interface, with a suffix of "v<N>", where <N> is the virtual device
1377 number. Previously those virtual devices were named as if completely
1380 The ninth and later NPAR virtual devices will be named following the
1381 scheme used for the first eight NPAR partitions. Previously those
1382 devices were not renamed and the kernel default (eth<n>) was used.
1384 "net_id" will also generate names for PCI devices where the PCI
1385 network controller device does not have an associated slot number
1386 itself, but one of its parents does. Previously those devices were
1387 not renamed and the kernel default (eth<n>) was used.
1389 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
1390 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
1391 the unit. So, it is expected that the default behavior of
1392 systemd-logind is not changed. However, if distribution packagers or
1393 administrators disabled or modified IPAddressDeny= setting by a
1394 drop-in config file, then it may be necessary to update the file to
1395 re-enable AF_INET and AF_INET6 to support network user name services,
1398 * When the RestrictNamespaces= unit property is specified multiple
1399 times, then the specified types are merged now. Previously, only the
1400 last assignment was used. So, if distribution packagers or
1401 administrators modified the setting by a drop-in config file, then it
1402 may be necessary to update the file.
1404 * When OnFailure= is used in combination with Restart= on a service
1405 unit, then the specified units will no longer be triggered on
1406 failures that result in restarting. Previously, the specified units
1407 would be activated each time the unit failed, even when the unit was
1408 going to be restarted automatically. This behaviour contradicted the
1409 documentation. With this release the code is adjusted to match the
1412 * systemd-tmpfiles will now print a notice whenever it encounters
1413 tmpfiles.d/ lines referencing the /var/run/ directory. It will
1414 recommend reworking them to use the /run/ directory instead (for
1415 which /var/run/ is simply a symlinked compatibility alias). This way
1416 systemd-tmpfiles can properly detect line conflicts and merge lines
1417 referencing the same file by two paths, without having to access
1420 * systemctl disable/unmask/preset/preset-all cannot be used with
1421 --runtime. Previously this was allowed, but resulted in unintuitive
1422 behaviour that wasn't useful. systemctl disable/unmask will now undo
1423 both runtime and persistent enablement/masking, i.e. it will remove
1424 any relevant symlinks both in /run and /etc.
1426 * Note that all long-running system services shipped with systemd will
1427 now default to a system call whitelist (rather than a blacklist, as
1428 before). In particular, systemd-udevd will now enforce one too. For
1429 most cases this should be safe, however downstream distributions
1430 which disabled sandboxing of systemd-udevd (specifically the
1431 MountFlags= setting), might want to disable this security feature
1432 too, as the default whitelisting will prohibit all mount, swap,
1433 reboot and clock changing operations from udev rules.
1435 * sd-boot acquired new loader configuration settings to optionally turn
1436 off Windows and MacOS boot partition discovery as well as
1437 reboot-into-firmware menu items. It is also able to pick a better
1438 screen resolution for HiDPI systems, and now provides loader
1439 configuration settings to change the resolution explicitly.
1441 * systemd-resolved now supports DNS-over-TLS. It's still
1442 turned off by default, use DNSOverTLS=opportunistic to turn it on in
1443 resolved.conf. We intend to make this the default as soon as couple
1444 of additional techniques for optimizing the initial latency caused by
1445 establishing a TLS/TCP connection are implemented.
1447 * systemd-resolved.service and systemd-networkd.service now set
1448 DynamicUser=yes. The users systemd-resolve and systemd-network are
1449 not created by systemd-sysusers anymore.
1451 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
1452 that embed a network facing module into any process using getpwuid()
1453 or related call: the dynamic allocation of the user ID for
1454 systemd-resolved.service means the service manager has to check NSS
1455 if the user name is already taken when forking off the service. Since
1456 the user in the common case won't be defined in /etc/passwd the
1457 lookup is likely to trigger nss-ldap which in turn might use NSS to
1458 ask systemd-resolved for hostname lookups. This will hence result in
1459 a deadlock: a user name lookup in order to start
1460 systemd-resolved.service will result in a host name lookup for which
1461 systemd-resolved.service needs to be started already. There are
1462 multiple ways to work around this problem: pre-allocate the
1463 "systemd-resolve" user on such systems, so that nss-ldap won't be
1464 triggered; or use a different NSS package that doesn't do networking
1465 in-process but provides a local asynchronous name cache; or configure
1466 the NSS package to avoid lookups for UIDs in the range `pkg-config
1467 systemd --variable=dynamicuidmin` … `pkg-config systemd
1468 --variable=dynamicuidmax`, so that it does not consider itself
1469 authoritative for the same UID range systemd allocates dynamic users
1472 * The systemd-resolve tool has been renamed to resolvectl (it also
1473 remains available under the old name, for compatibility), and its
1474 interface is now verb-based, similar in style to the other <xyz>ctl
1475 tools, such as systemctl or loginctl.
1477 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
1478 compatibility. It may be symlinked under the 'resolvconf' name, in
1479 which case it will take arguments and input compatible with the
1480 Debian and FreeBSD resolvconf tool.
1482 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
1483 where the system initially suspends, and after a timeout resumes and
1486 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
1487 set the client will only send a DUID as client identifier.
1489 * The nss-systemd glibc NSS module will now enumerate dynamic users and
1490 groups in effect. Previously, it could resolve UIDs/GIDs to user
1491 names/groups and vice versa, but did not support enumeration.
1493 * journald's Compress= configuration setting now optionally accepts a
1494 byte threshold value. All journal objects larger than this threshold
1495 will be compressed, smaller ones will not. Previously this threshold
1496 was not configurable and set to 512.
1498 * A new system.conf setting NoNewPrivileges= is now available which may
1499 be used to turn off acquisition of new privileges system-wide
1500 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
1501 for all its children). Note that turning this option on means setuid
1502 binaries and file system capabilities lose their special powers.
1503 While turning on this option is a big step towards a more secure
1504 system, doing so is likely to break numerous pre-existing UNIX tools,
1505 in particular su and sudo.
1507 * A new service systemd-time-sync-wait.service has been added. If
1508 enabled it will delay the time-sync.target unit at boot until time
1509 synchronization has been received from the network. This
1510 functionality is useful on systems lacking a local RTC or where it is
1511 acceptable that the boot process shall be delayed by external network
1514 * When hibernating, systemd will now inform the kernel of the image
1515 write offset, on kernels new enough to support this. This means swap
1516 files should work for hibernation now.
1518 * When loading unit files, systemd will now look for drop-in unit files
1519 extensions in additional places. Previously, for a unit file name
1520 "foo-bar-baz.service" it would look for dropin files in
1521 "foo-bar-baz.service.d/*.conf". Now, it will also look in
1522 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
1523 service name truncated after all inner dashes. This scheme allows
1524 writing drop-ins easily that apply to a whole set of unit files at
1525 once. It's particularly useful for mount and slice units (as their
1526 naming is prefix based), but is also useful for service and other
1527 units, for packages that install multiple unit files at once,
1528 following a strict naming regime of beginning the unit file name with
1529 the package's name. Two new specifiers are now supported in unit
1530 files to match this: %j and %J are replaced by the part of the unit
1531 name following the last dash.
1533 * Unit files and other configuration files that support specifier
1534 expansion now understand another three new specifiers: %T and %V will
1535 resolve to /tmp and /var/tmp respectively, or whatever temporary
1536 directory has been set for the calling user. %E will expand to either
1537 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
1539 * The ExecStart= lines of unit files are no longer required to
1540 reference absolute paths. If non-absolute paths are specified the
1541 specified binary name is searched within the service manager's
1542 built-in $PATH, which may be queried with 'systemd-path
1543 search-binaries-default'. It's generally recommended to continue to
1544 use absolute paths for all binaries specified in unit files.
1546 * Units gained a new load state "bad-setting", which is used when a
1547 unit file was loaded, but contained fatal errors which prevent it
1548 from being started (for example, a service unit has been defined
1549 lacking both ExecStart= and ExecStop= lines).
1551 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
1552 support alternative debuggers, for example lldb. The old name
1553 continues to be available however, for compatibility reasons. Use the
1554 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
1555 to pick an alternative debugger instead of the default gdb.
1557 * systemctl and the other tools will now output escape sequences that
1558 generate proper clickable hyperlinks in various terminal emulators
1559 where useful (for example, in the "systemctl status" output you can
1560 now click on the unit file name to quickly open it in the
1561 editor/viewer of your choice). Note that not all terminal emulators
1562 support this functionality yet, but many do. Unfortunately, the
1563 "less" pager doesn't support this yet, hence this functionality is
1564 currently automatically turned off when a pager is started (which
1565 happens quite often due to auto-paging). We hope to remove this
1566 limitation as soon as "less" learns these escape sequences. This new
1567 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
1568 environment variable. For details on these escape sequences see:
1569 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
1571 * networkd's .network files now support a new IPv6MTUBytes= option for
1572 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
1573 option in the [Route] section to configure the MTU to use for
1574 specific routes. It also gained support for configuration of the DHCP
1575 "UserClass" option through the new UserClass= setting. It gained
1576 three new options in the new [CAN] section for configuring CAN
1577 networks. The MULTICAST and ALLMULTI interface flags may now be
1578 controlled explicitly with the new Multicast= and AllMulticast=
1581 * networkd will now automatically make use of the kernel's route
1582 expiration feature, if it is available.
1584 * udevd's .link files now support setting the number of receive and
1585 transmit channels, using the RxChannels=, TxChannels=,
1586 OtherChannels=, CombinedChannels= settings.
1588 * Support for UDPSegmentationOffload= has been removed, given its
1589 limited support in hardware, and waning software support.
1591 * networkd's .netdev files now support creating "netdevsim" interfaces.
1593 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
1594 to query the unit belonging to a specific kernel control group.
1596 * systemd-analyze gained a new verb "cat-config", which may be used to
1597 dump the contents of any configuration file, with all its matching
1598 drop-in files added in, and honouring the usual search and masking
1599 logic applied to systemd configuration files. For example use
1600 "systemd-analyze cat-config systemd/system.conf" to get the complete
1601 system configuration file of systemd how it would be loaded by PID 1
1602 itself. Similar to this, various tools such as systemd-tmpfiles or
1603 systemd-sysusers, gained a new option "--cat-config", which does the
1604 corresponding operation for their own configuration settings. For
1605 example, "systemd-tmpfiles --cat-config" will now output the full
1606 list of tmpfiles.d/ lines in place.
1608 * timedatectl gained three new verbs: "show" shows bus properties of
1609 systemd-timedated, "timesync-status" shows the current NTP
1610 synchronization state of systemd-timesyncd, and "show-timesync"
1611 shows bus properties of systemd-timesyncd.
1613 * systemd-timesyncd gained a bus interface on which it exposes details
1616 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
1617 understood by systemd-timedated. It takes a colon-separated list of
1618 unit names of NTP client services. The list is used by
1619 "timedatectl set-ntp".
1621 * systemd-nspawn gained a new --rlimit= switch for setting initial
1622 resource limits for the container payload. There's a new switch
1623 --hostname= to explicitly override the container's hostname. A new
1624 --no-new-privileges= switch may be used to control the
1625 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
1626 --oom-score-adjust= switch controls the OOM scoring adjustment value
1627 for the payload. The new --cpu-affinity= switch controls the CPU
1628 affinity of the container payload. The new --resolv-conf= switch
1629 allows more detailed control of /etc/resolv.conf handling of the
1630 container. Similarly, the new --timezone= switch allows more detailed
1631 control of /etc/localtime handling of the container.
1633 * systemd-detect-virt gained a new --list switch, which will print a
1634 list of all currently known VM and container environments.
1636 * Support for "Portable Services" has been added, see
1637 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
1638 experimental, but this is expected to change soon. Reflecting this
1639 experimental state, the "portablectl" binary is not installed into
1640 /usr/bin yet. The binary has to be called with the full path
1641 /usr/lib/systemd/portablectl instead.
1643 * journalctl's and systemctl's -o switch now knows a new log output
1644 mode "with-unit". The output it generates is very similar to the
1645 regular "short" mode, but displays the unit name instead of the
1646 syslog tag for each log line. Also, the date is shown with timezone
1647 information. This mode is probably more useful than the classic
1648 "short" output mode for most purposes, except where pixel-perfect
1649 compatibility with classic /var/log/messages formatting is required.
1651 * A new --dump-bus-properties switch has been added to the systemd
1652 binary, which may be used to dump all supported D-Bus properties.
1653 (Options which are still supported, but are deprecated, are *not*
1656 * sd-bus gained a set of new calls:
1657 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
1658 enable/disable the "floating" state of a bus slot object,
1659 i.e. whether the slot object pins the bus it is allocated for into
1660 memory or if the bus slot object gets disconnected when the bus goes
1661 away. sd_bus_open_with_description(),
1662 sd_bus_open_user_with_description(),
1663 sd_bus_open_system_with_description() may be used to allocate bus
1664 objects and set their description string already during allocation.
1666 * sd-event gained support for watching inotify events from the event
1667 loop, in an efficient way, sharing inotify handles between multiple
1668 users. For this a new function sd_event_add_inotify() has been added.
1670 * sd-event and sd-bus gained support for calling special user-supplied
1671 destructor functions for userdata pointers associated with
1672 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
1673 functions sd_bus_slot_set_destroy_callback,
1674 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
1675 sd_bus_track_get_destroy_callback,
1676 sd_event_source_set_destroy_callback,
1677 sd_event_source_get_destroy_callback have been added.
1679 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
1681 * PID 1 will now automatically reschedule .timer units whenever the
1682 local timezone changes. (They previously got rescheduled
1683 automatically when the system clock changed.)
1685 * New documentation has been added to document cgroups delegation,
1686 portable services and the various code quality tools we have set up:
1688 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
1689 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
1690 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
1692 * The Boot Loader Specification has been added to the source tree.
1694 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
1696 While moving it into our source tree we have updated it and further
1697 changes are now accepted through the usual github PR workflow.
1699 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
1700 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
1701 earlier PAM modules. The data in these fields is used to initialize
1702 the session scope's resource properties. Thus external PAM modules
1703 may now configure per-session limits, for example sourced from
1704 external user databases.
1706 * socket units with Accept=yes will now maintain a "refused" counter in
1707 addition to the existing "accepted" counter, counting connections
1708 refused due to the enforced limits.
1710 * The "systemd-path search-binaries-default" command may now be use to
1711 query the default, built-in $PATH PID 1 will pass to the services it
1714 * A new unit file setting PrivateMounts= has been added. It's a boolean
1715 option. If enabled the unit's processes are invoked in their own file
1716 system namespace. Note that this behaviour is also implied if any
1717 other file system namespacing options (such as PrivateTmp=,
1718 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
1719 primarily useful for services that do not use any of the other file
1720 system namespacing options. One such service is systemd-udevd.service
1721 where this is now used by default.
1723 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
1724 when the system is booted in UEFI "secure mode".
1726 * A new unit "system-update-pre.target" is added, which defines an
1727 optional synchronization point for offline system updates, as
1728 implemented by the pre-existing "system-update.target" unit. It
1729 allows ordering services before the service that executes the actual
1730 update process in a generic way.
1732 * Systemd now emits warnings whenever .include syntax is used.
1734 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
1735 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
1736 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
1737 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
1738 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
1739 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
1740 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
1741 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
1742 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
1743 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
1744 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
1745 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
1746 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
1747 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
1748 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
1749 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
1750 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
1751 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
1752 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
1753 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
1754 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
1755 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
1756 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
1757 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
1758 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
1759 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
1760 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
1761 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
1762 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1764 — Berlin, 2018-06-22
1768 * The MemoryAccounting= unit property now defaults to on. After
1769 discussions with the upstream control group maintainers we learnt
1770 that the negative impact of cgroup memory accounting on current
1771 kernels is finally relatively minimal, so that it should be safe to
1772 enable this by default without affecting system performance. Besides
1773 memory accounting only task accounting is turned on by default, all
1774 other forms of resource accounting (CPU, IO, IP) remain off for now,
1775 because it's not clear yet that their impact is small enough to move
1776 from opt-in to opt-out. We recommend downstreams to leave memory
1777 accounting on by default if kernel 4.14 or higher is primarily
1778 used. On very resource constrained systems or when support for old
1779 kernels is a necessity, -Dmemory-accounting-default=false can be used
1780 to revert this change.
1782 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
1783 %udev_rules_update) and the journal catalog (%journal_catalog_update)
1784 from the upgrade scriptlets of individual packages now do nothing.
1785 Transfiletriggers have been added which will perform those updates
1786 once at the end of the transaction.
1788 Similar transfiletriggers have been added to execute any sysctl.d
1789 and binfmt.d rules. Thus, it should be unnecessary to provide any
1790 scriptlets to execute this configuration from package installation
1793 * systemd-sysusers gained a mode where the configuration to execute is
1794 specified on the command line, but this configuration is not executed
1795 directly, but instead it is merged with the configuration on disk,
1796 and the result is executed. This is useful for package installation
1797 scripts which want to create the user before installing any files on
1798 disk (in case some of those files are owned by that user), while
1799 still allowing local admin overrides.
1801 This functionality is exposed to rpm scriptlets through a new
1802 %sysusers_create_package macro. Old %sysusers_create and
1803 %sysusers_create_inline macros are deprecated.
1805 A transfiletrigger for sysusers.d configuration is now installed,
1806 which means that it should be unnecessary to call systemd-sysusers from
1807 package installation scripts, unless the package installs any files
1808 owned by those newly-created users, in which case
1809 %sysusers_create_package should be used.
1811 * Analogous change has been done for systemd-tmpfiles: it gained a mode
1812 where the command-line configuration is merged with the configuration
1813 on disk. This is exposed as the new %tmpfiles_create_package macro,
1814 and %tmpfiles_create is deprecated. A transfiletrigger is installed
1815 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
1816 from package installation scripts.
1818 * sysusers.d configuration for a user may now also specify the group
1819 number, in addition to the user number ("u username 123:456"), or
1820 without the user number ("u username -:456").
1822 * Configution items for systemd-sysusers can now be specified as
1823 positional arguments when the new --inline switch is used.
1825 * The login shell of users created through sysusers.d may now be
1826 specified (previously, it was always /bin/sh for root and
1827 /sbin/nologin for other users).
1829 * systemd-analyze gained a new --global switch to look at global user
1830 configuration. It also gained a unit-paths verb to list the unit load
1831 paths that are compiled into systemd (which can be used with
1832 --systemd, --user, or --global).
1834 * udevadm trigger gained a new --settle/-w option to wait for any
1835 triggered events to finish (but just those, and not any other events
1836 which are triggered meanwhile).
1838 * The action that systemd-logind takes when the lid is closed and the
1839 machine is connected to external power can now be configured using
1840 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
1841 was determined by HandleLidSwitch=, and, for backwards compatibility,
1842 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
1844 * journalctl will periodically call sd_journal_process() to make it
1845 resilient against inotify queue overruns when journal files are
1846 rotated very quickly.
1848 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
1849 sd_bus_get_n_queued_write — may be used to check the number of
1850 pending bus messages.
1852 * systemd gained a new
1853 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
1854 which can be used to migrate foreign processes to scope and service
1855 units. The primary user for this new API is systemd itself: the
1856 systemd --user instance uses this call of the systemd --system
1857 instance to migrate processes if it itself gets the request to
1858 migrate processes and the kernel refuses this due to access
1859 restrictions. Thanks to this "systemd-run --scope --user …" works
1860 again in pure cgroup v2 environments when invoked from the user
1863 * A new TemporaryFileSystem= setting can be used to mask out part of
1864 the real file system tree with tmpfs mounts. This may be combined
1865 with BindPaths= and BindReadOnlyPaths= to hide files or directories
1866 not relevant to the unit, while still allowing some paths lower in
1867 the tree to be accessed.
1869 ProtectHome=tmpfs may now be used to hide user home and runtime
1870 directories from units, in a way that is mostly equivalent to
1871 "TemporaryFileSystem=/home /run/user /root".
1873 * Non-service units are now started with KeyringMode=shared by default.
1874 This means that mount and swapon and other mount tools have access
1875 to keys in the main keyring.
1877 * /sys/fs/bpf is now mounted automatically.
1879 * QNX virtualization is now detected by systemd-detect-virt and may
1880 be used in ConditionVirtualization=.
1882 * IPAccounting= may now be enabled also for slice units.
1884 * A new -Dsplit-bin= build configuration switch may be used to specify
1885 whether bin and sbin directories are merged, or if they should be
1886 included separately in $PATH and various listings of executable
1887 directories. The build configuration scripts will try to autodetect
1888 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
1889 system, but distributions are encouraged to configure this
1892 * A new -Dok-color= build configuration switch may be used to change
1893 the colour of "OK" status messages.
1895 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
1896 PrivateNetwork=yes was buggy in previous versions of systemd. This
1897 means that after the upgrade and daemon-reexec, any such units must
1900 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
1901 will not exclude read-only files owned by root from cleanup.
1903 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
1904 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
1905 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
1906 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
1907 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
1908 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
1909 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
1910 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
1911 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
1912 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
1913 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
1914 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
1915 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1916 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
1917 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
1918 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
1920 — Warsaw, 2018-03-05
1924 * Some keyboards come with a zoom see-saw or rocker which until now got
1925 mapped to the Linux "zoomin/out" keys in hwdb. However, these
1926 keycodes are not recognized by any major desktop. They now produce
1927 Up/Down key events so that they can be used for scrolling.
1929 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
1930 slightly: previously, if an argument was specified for lines of this
1931 type (i.e. the right-most column was set) this string was appended to
1932 existing files each time systemd-tmpfiles was run. This behaviour was
1933 different from what the documentation said, and not particularly
1934 useful, as repeated systemd-tmpfiles invocations would not be
1935 idempotent and grow such files without bounds. With this release
1936 behaviour has been altered to match what the documentation says:
1937 lines of this type only have an effect if the indicated files don't
1938 exist yet, and only then the argument string is written to the file.
1940 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
1941 systemd-tmpfiles behaviour: previously, read-only files owned by root
1942 were always excluded from the file "aging" algorithm (i.e. the
1943 automatic clean-up of directories like /tmp based on
1944 atime/mtime/ctime). We intend to drop this restriction, and age files
1945 by default even when owned by root and read-only. This behaviour was
1946 inherited from older tools, but there have been requests to remove
1947 it, and it's not obvious why this restriction was made in the first
1948 place. Please speak up now, if you are aware of software that reqires
1949 this behaviour, otherwise we'll remove the restriction in v238.
1951 * A new environment variable $SYSTEMD_OFFLINE is now understood by
1952 systemctl. It takes a boolean argument. If on, systemctl assumes it
1953 operates on an "offline" OS tree, and will not attempt to talk to the
1954 service manager. Previously, this mode was implicitly enabled if a
1955 chroot() environment was detected, and this new environment variable
1956 now provides explicit control.
1958 * .path and .socket units may now be created transiently, too.
1959 Previously only service, mount, automount and timer units were
1960 supported as transient units. The systemd-run tool has been updated
1961 to expose this new functionality, you may hence use it now to bind
1962 arbitrary commands to path or socket activation on-the-fly from the
1963 command line. Moreover, almost all properties are now exposed for the
1964 unit types that already supported transient operation.
1966 * The systemd-mount command gained support for a new --owner= parameter
1967 which takes a user name, which is then resolved and included in uid=
1968 and gid= mount options string of the file system to mount.
1970 * A new unit condition ConditionControlGroupController= has been added
1971 that checks whether a specific cgroup controller is available.
1973 * Unit files, udev's .link files, and systemd-networkd's .netdev and
1974 .network files all gained support for a new condition
1975 ConditionKernelVersion= for checking against specific kernel
1978 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
1979 support for configuring device flags in the Flags= setting. In the
1980 same files, the [Tunnel] section gained support for configuring
1981 AllowLocalRemote=. The [Route] section in .network files gained
1982 support for configuring InitialCongestionWindow=,
1983 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
1984 understands RapidCommit=.
1986 * systemd-networkd's DHCPv6 support gained support for Prefix
1989 * sd-bus gained support for a new "watch-bind" feature. When this
1990 feature is enabled, an sd_bus connection may be set up to connect to
1991 an AF_UNIX socket in the file system as soon as it is created. This
1992 functionality is useful for writing early-boot services that
1993 automatically connect to the system bus as soon as it is started,
1994 without ugly time-based polling. systemd-networkd and
1995 systemd-resolved have been updated to make use of this
1996 functionality. busctl exposes this functionality in a new
1997 --watch-bind= command line switch.
1999 * sd-bus will now optionally synthesize a local "Connected" signal as
2000 soon as a D-Bus connection is set up fully. This message mirrors the
2001 already existing "Disconnected" signal which is synthesized when the
2002 connection is terminated. This signal is generally useful but
2003 particularly handy in combination with the "watch-bind" feature
2004 described above. Synthesizing of this message has to be requested
2005 explicitly through the new API call sd_bus_set_connected_signal(). In
2006 addition a new call sd_bus_is_ready() has been added that checks
2007 whether a connection is fully set up (i.e. between the "Connected" and
2008 "Disconnected" signals).
2010 * sd-bus gained two new calls sd_bus_request_name_async() and
2011 sd_bus_release_name_async() for asynchronously registering bus
2012 names. Similar, there is now sd_bus_add_match_async() for installing
2013 a signal match asynchronously. All of systemd's own services have
2014 been updated to make use of these calls. Doing these operations
2015 asynchronously has two benefits: it reduces the risk of deadlocks in
2016 case of cyclic dependencies between bus services, and it speeds up
2017 service initialization since synchronization points for bus
2018 round-trips are removed.
2020 * sd-bus gained two new calls sd_bus_match_signal() and
2021 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
2022 and sd_bus_add_match_async() but instead of taking a D-Bus match
2023 string take match fields as normal function parameters.
2025 * sd-bus gained two new calls sd_bus_set_sender() and
2026 sd_bus_message_set_sender() for setting the sender name of outgoing
2027 messages (either for all outgoing messages or for just one specific
2028 one). These calls are only useful in direct connections as on
2029 brokered connections the broker fills in the sender anyway,
2030 overwriting whatever the client filled in.
2032 * sd-event gained a new pseudo-handle that may be specified on all API
2033 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
2034 used this refers to the default event loop object of the calling
2035 thread. Note however that this does not implicitly allocate one —
2036 which has to be done prior by using sd_event_default(). Similarly
2037 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
2038 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
2039 to the default bus of the specified type of the calling thread. Here
2040 too this does not implicitly allocate bus connection objects, this
2041 has to be done prior with sd_bus_default() and friends.
2043 * sd-event gained a new call pair
2044 sd_event_source_{get|set}_io_fd_own(). This may be used to request
2045 automatic closure of the file descriptor an IO event source watches
2046 when the event source is destroyed.
2048 * systemd-networkd gained support for natively configuring WireGuard
2051 * In previous versions systemd synthesized user records both for the
2052 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
2053 internally. In order to simplify distribution-wide renames of the
2054 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
2055 new transitional flag file has been added: if
2056 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
2057 user and group record within the systemd codebase is disabled.
2059 * systemd-notify gained a new --uid= option for selecting the source
2060 user/UID to use for notification messages sent to the service
2063 * journalctl gained a new --grep= option to list only entries in which
2064 the message matches a certain pattern. By default matching is case
2065 insensitive if the pattern is lowercase, and case sensitive
2066 otherwise. Option --case-sensitive=yes|no can be used to override
2067 this an specify case sensitivity or case insensitivity.
2069 * There's now a "systemd-analyze service-watchdogs" command for printing
2070 the current state of the service runtime watchdog, and optionally
2071 enabling or disabling the per-service watchdogs system-wide if given a
2072 boolean argument (i.e. the concept you configure in WatchdogSec=), for
2073 debugging purposes. There's also a kernel command line option
2074 systemd.service_watchdogs= for controlling the same.
2076 * Two new "log-level" and "log-target" options for systemd-analyze were
2077 added that merge the now deprecated get-log-level, set-log-level and
2078 get-log-target, set-log-target pairs. The deprecated options are still
2079 understood for backwards compatibility. The two new options print the
2080 current value when no arguments are given, and set them when a
2081 level/target is given as an argument.
2083 * sysusers.d's "u" lines now optionally accept both a UID and a GID
2084 specification, separated by a ":" character, in order to create users
2085 where UID and GID do not match.
2087 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
2088 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
2089 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
2090 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
2091 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
2092 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
2093 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
2094 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
2095 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
2096 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
2097 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
2098 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
2099 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
2100 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
2101 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
2102 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
2103 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
2104 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
2105 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
2106 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
2113 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
2114 in v235 has been extended to also set the dummy.ko module option
2115 numdummies=0, preventing the kernel from automatically creating
2116 dummy0. All dummy interfaces must now be explicitly created.
2118 * Unknown '%' specifiers in configuration files are now rejected. This
2119 applies to units and tmpfiles.d configuration. Any percent characters
2120 that are followed by a letter or digit that are not supposed to be
2121 interpreted as the beginning of a specifier should be escaped by
2122 doubling ("%%"). (So "size=5%" is still accepted, as well as
2123 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
2124 valid specifiers today.)
2126 * systemd-resolved now maintains a new dynamic
2127 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
2128 recommended to make /etc/resolv.conf a symlink to it. This file
2129 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
2130 includes dynamically acquired search domains, achieving more correct
2131 DNS resolution by software that bypasses local DNS APIs such as NSS.
2133 * The "uaccess" udev tag has been dropped from /dev/kvm and
2134 /dev/dri/renderD*. These devices now have the 0666 permissions by
2135 default (but this may be changed at build-time). /dev/dri/renderD*
2136 will now be owned by the "render" group along with /dev/kfd.
2138 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
2139 systemd-journal-gatewayd.service and
2140 systemd-journal-upload.service. This means "nss-systemd" must be
2141 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
2142 services are resolved properly.
2144 * In /etc/fstab two new mount options are now understood:
2145 x-systemd.makefs and x-systemd.growfs. The former has the effect that
2146 the configured file system is formatted before it is mounted, the
2147 latter that the file system is resized to the full block device size
2148 after it is mounted (i.e. if the file system is smaller than the
2149 partition it resides on, it's grown). This is similar to the fsck
2150 logic in /etc/fstab, and pulls in systemd-makefs@.service and
2151 systemd-growfs@.service as necessary, similar to
2152 systemd-fsck@.service. Resizing is currently only supported on ext4
2155 * In systemd-networkd, the IPv6 RA logic now optionally may announce
2156 DNS server and domain information.
2158 * Support for the LUKS2 on-disk format for encrypted partitions has
2159 been added. This requires libcryptsetup2 during compilation and
2162 * The systemd --user instance will now signal "readiness" when its
2163 basic.target unit has been reached, instead of when the run queue ran
2164 empty for the first time.
2166 * Tmpfiles.d with user configuration are now also supported.
2167 systemd-tmpfiles gained a new --user switch, and snippets placed in
2168 ~/.config/user-tmpfiles.d/ and corresponding directories will be
2169 executed by systemd-tmpfiles --user running in the new
2170 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
2171 running in the user session.
2173 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
2174 %S resolves to the top-level state directory (/var/lib for the system
2175 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
2176 top-level cache directory (/var/cache for the system instance,
2177 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
2178 logs directory (/var/log for the system instance,
2179 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
2180 existing %t specifier, that resolves to the top-level runtime
2181 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
2184 * journalctl learnt a new parameter --output-fields= for limiting the
2185 set of journal fields to output in verbose and JSON output modes.
2187 * systemd-timesyncd's configuration file gained a new option
2188 RootDistanceMaxSec= for setting the maximum root distance of servers
2189 it'll use, as well as the new options PollIntervalMinSec= and
2190 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
2192 * bootctl gained a new command "list" for listing all available boot
2193 menu items on systems that follow the boot loader specification.
2195 * systemctl gained a new --dry-run switch that shows what would be done
2196 instead of doing it, and is currently supported by the shutdown and
2199 * ConditionSecurity= can now detect the TOMOYO security module.
2201 * Unit file [Install] sections are now also respected in unit drop-in
2202 files. This is intended to be used by drop-ins under /usr/lib/.
2204 * systemd-firstboot may now also set the initial keyboard mapping.
2206 * Udev "changed" events for devices which are exposed as systemd
2207 .device units are now propagated to units specified in
2208 ReloadPropagatedFrom= as reload requests.
2210 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
2211 unit template name (i.e. a name in the form of 'foobar@.service',
2212 without the instance component between the '@' and - the '.'), then
2213 the escaped sysfs path of the device is automatically used as the
2216 * SystemCallFilter= in unit files has been extended so that an "errno"
2217 can be specified individually for each system call. Example:
2218 SystemCallFilter=~uname:EILSEQ.
2220 * The cgroup delegation logic has been substantially updated. Delegate=
2221 now optionally takes a list of controllers (instead of a boolean, as
2222 before), which lists the controllers to delegate at least.
2224 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
2226 * A new LogLevelMax= setting configures the maximum log level any
2227 process of the service may log at (i.e. anything with a lesser
2228 priority than what is specified is automatically dropped). A new
2229 LogExtraFields= setting allows configuration of additional journal
2230 fields to attach to all log records generated by any of the unit's
2233 * New StandardInputData= and StandardInputText= settings along with the
2234 new option StandardInput=data may be used to configure textual or
2235 binary data that shall be passed to the executed service process via
2236 standard input, encoded in-line in the unit file.
2238 * StandardInput=, StandardOutput= and StandardError= may now be used to
2239 connect stdin/stdout/stderr of executed processes directly with a
2240 file or AF_UNIX socket in the file system, using the new "file:" option.
2242 * A new unit file option CollectMode= has been added, that allows
2243 tweaking the garbage collection logic for units. It may be used to
2244 tell systemd to garbage collect units that have failed automatically
2245 (normally it only GCs units that exited successfully). systemd-run
2246 and systemd-mount expose this new functionality with a new -G option.
2248 * "machinectl bind" may now be used to bind mount non-directories
2249 (i.e. regularfiles, devices, fifos, sockets).
2251 * systemd-analyze gained a new verb "calendar" for validating and
2252 testing calendar time specifications to use for OnCalendar= in timer
2253 units. Besides validating the expression it will calculate the next
2254 time the specified expression would elapse.
2256 * In addition to the pre-existing FailureAction= unit file setting
2257 there's now SuccessAction=, for configuring a shutdown action to
2258 execute when a unit completes successfully. This is useful in
2259 particular inside containers that shall terminate after some workload
2260 has been completed. Also, both options are now supported for all unit
2261 types, not just services.
2263 * networkds's IP rule support gained two new options
2264 IncomingInterface= and OutgoingInterface= for configuring the incoming
2265 and outgoing interfaces of configured rules. systemd-networkd also
2266 gained support for "vxcan" network devices.
2268 * networkd gained a new setting RequiredForOnline=, taking a
2269 boolean. If set, systemd-wait-online will take it into consideration
2270 when determining that the system is up, otherwise it will ignore the
2271 interface for this purpose.
2273 * The sd_notify() protocol gained support for a new operation: with
2274 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
2275 store again, ahead of POLLHUP or POLLERR when they are removed
2278 * A new document doc/UIDS-GIDS.md has been added to the source tree,
2279 that documents the UID/GID range and assignment assumptions and
2280 requirements of systemd.
2282 * The watchdog device PID 1 will ping may now be configured through the
2283 WatchdogDevice= configuration file setting, or by setting the
2284 systemd.watchdog_service= kernel commandline option.
2286 * systemd-resolved's gained support for registering DNS-SD services on
2287 the local network using MulticastDNS. Services may either be
2288 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
2289 the same dir below /run, /usr/lib), or through its D-Bus API.
2291 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
2292 extend the effective start, runtime, and stop time. The service must
2293 continue to send EXTEND_TIMEOUT_USEC within the period specified to
2294 prevent the service manager from making the service as timedout.
2296 * systemd-resolved's DNSSEC support gained support for RFC 8080
2297 (Ed25519 keys and signatures).
2299 * The systemd-resolve command line tool gained a new set of options
2300 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
2301 --set-nta= and --revert to configure per-interface DNS configuration
2302 dynamically during runtime. It's useful for pushing DNS information
2303 into systemd-resolved from DNS hook scripts that various interface
2304 managing software supports (such as pppd).
2306 * systemd-nspawn gained a new --network-namespace-path= command line
2307 option, which may be used to make a container join an existing
2308 network namespace, by specifying a path to a "netns" file.
2310 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
2311 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
2312 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
2313 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
2314 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
2315 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
2316 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
2317 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
2318 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
2319 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
2320 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
2321 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
2322 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
2323 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
2324 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
2325 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
2326 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
2327 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
2328 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
2329 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
2330 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
2331 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
2332 Jędrzejewski-Szmek, Zeal Jagannatha
2334 — Berlin, 2017-12-14
2338 * INCOMPATIBILITY: systemd-logind.service and other long-running
2339 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
2340 communication with the outside. This generally improves security of
2341 the system, and is in almost all cases a safe and good choice, as
2342 these services do not and should not provide any network-facing
2343 functionality. However, systemd-logind uses the glibc NSS API to
2344 query the user database. This creates problems on systems where NSS
2345 is set up to directly consult network services for user database
2346 lookups. In particular, this creates incompatibilities with the
2347 "nss-nis" module, which attempts to directly contact the NIS/YP
2348 network servers it is configured for, and will now consistently
2349 fail. In such cases, it is possible to turn off IP sandboxing for
2350 systemd-logind.service (set IPAddressDeny= in its [Service] section
2351 to the empty string, via a .d/ unit file drop-in). Downstream
2352 distributions might want to update their nss-nis packaging to include
2353 such a drop-in snippet, accordingly, to hide this incompatibility
2354 from the user. Another option is to make use of glibc's nscd service
2355 to proxy such network requests through a privilege-separated, minimal
2356 local caching daemon, or to switch to more modern technologies such
2357 sssd, whose NSS hook-ups generally do not involve direct network
2358 access. In general, we think it's definitely time to question the
2359 implementation choices of nss-nis, i.e. whether it's a good idea
2360 today to embed a network-facing loadable module into all local
2361 processes that need to query the user database, including the most
2362 trivial and benign ones, such as "ls". For more details about
2363 IPAddressDeny= see below.
2365 * A new modprobe.d drop-in is now shipped by default that sets the
2366 bonding module option max_bonds=0. This overrides the kernel default,
2367 to avoid conflicts and ambiguity as to whether or not bond0 should be
2368 managed by systemd-networkd or not. This resolves multiple issues
2369 with bond0 properties not being applied, when bond0 is configured
2370 with systemd-networkd. Distributors may choose to not package this,
2371 however in that case users will be prevented from correctly managing
2372 bond0 interface using systemd-networkd.
2374 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
2375 which print the logging level and target of the system manager. They
2376 complement the existing "set-log-level" and "set-log-target" verbs
2377 used to change those values.
2379 * journald.conf gained a new boolean setting ReadKMsg= which defaults
2380 to on. If turned off kernel log messages will not be read by
2381 systemd-journald or included in the logs. It also gained a new
2382 setting LineMax= for configuring the maximum line length in
2383 STDOUT/STDERR log streams. The new default for this value is 48K, up
2384 from the previous hardcoded 2048.
2386 * A new unit setting RuntimeDirectoryPreserve= has been added, which
2387 allows more detailed control of what to do with a runtime directory
2388 configured with RuntimeDirectory= (i.e. a directory below /run or
2389 $XDG_RUNTIME_DIR) after a unit is stopped.
2391 * The RuntimeDirectory= setting for units gained support for creating
2392 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
2393 one top-level directory.
2395 * Units gained new options StateDirectory=, CacheDirectory=,
2396 LogsDirectory= and ConfigurationDirectory= which are closely related
2397 to RuntimeDirectory= but manage per-service directories below
2398 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
2399 possible to write unit files which when activated automatically gain
2400 properly owned service specific directories in these locations, thus
2401 making unit files self-contained and increasing compatibility with
2402 stateless systems and factory reset where /etc or /var are
2403 unpopulated at boot. Matching these new settings there's also
2404 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
2405 ConfigurationDirectoryMode= for configuring the access mode of these
2406 directories. These settings are particularly useful in combination
2407 with DynamicUser=yes as they provide secure, properly-owned,
2408 writable, and stateful locations for storage, excluded from the
2409 sandbox that such services live in otherwise.
2411 * Automake support has been removed from this release. systemd is now
2414 * systemd-journald will now aggressively cache client metadata during
2415 runtime, speeding up log write performance under pressure. This comes
2416 at a small price though: as much of the metadata is read
2417 asynchronously from /proc/ (and isn't implicitly attached to log
2418 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
2419 metadata stored alongside a log entry might be slightly
2420 out-of-date. Previously it could only be slightly newer than the log
2421 message. The time window is small however, and given that the kernel
2422 is unlikely to be improved anytime soon in this regard, this appears
2425 * nss-myhostname/systemd-resolved will now by default synthesize an
2426 A/AAAA resource record for the "_gateway" hostname, pointing to the
2427 current default IP gateway. Previously it did that for the "gateway"
2428 name, hampering adoption, as some distributions wanted to leave that
2429 host name open for local use. The old behaviour may still be
2430 requested at build time.
2432 * systemd-networkd's [Address] section in .network files gained a new
2433 Scope= setting for configuring the IP address scope. The [Network]
2434 section gained a new boolean setting ConfigureWithoutCarrier= that
2435 tells systemd-networkd to ignore link sensing when configuring the
2436 device. The [DHCP] section gained a new Anonymize= boolean option for
2437 turning on a number of options suggested in RFC 7844. A new
2438 [RoutingPolicyRule] section has been added for configuring the IP
2439 routing policy. The [Route] section has gained support for a new
2440 Type= setting which permits configuring
2441 blackhole/unreachable/prohibit routes.
2443 * The [VRF] section in .netdev files gained a new Table= setting for
2444 configuring the routing table to use. The [Tunnel] section gained a
2445 new Independent= boolean field for configuring tunnels independent of
2446 an underlying network interface. The [Bridge] section gained a new
2447 GroupForwardMask= option for configuration of propagation of link
2448 local frames between bridge ports.
2450 * The WakeOnLan= setting in .link files gained support for a number of
2451 new modes. A new TCP6SegmentationOffload= setting has been added for
2452 configuring TCP/IPv6 hardware segmentation offload.
2454 * The IPv6 RA sender implementation may now optionally send out RDNSS
2455 and RDNSSL records to supply DNS configuration to peers.
2457 * systemd-nspawn gained support for a new --system-call-filter= command
2458 line option for adding and removing entries in the default system
2459 call filter it applies. Moreover systemd-nspawn has been changed to
2460 implement a system call whitelist instead of a blacklist.
2462 * systemd-run gained support for a new --pipe command line option. If
2463 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
2464 are directly passed on to the activated transient service
2465 executable. This allows invoking arbitrary processes as systemd
2466 services (for example to take benefit of dependency management,
2467 accounting management, resource management or log management that is
2468 done automatically for services) — while still allowing them to be
2469 integrated in a classic UNIX shell pipeline.
2471 * When a service sends RELOAD=1 via sd_notify() and reload propagation
2472 using ReloadPropagationTo= is configured, a reload is now propagated
2473 to configured units. (Previously this was only done on explicitly
2474 requested reloads, using "systemctl reload" or an equivalent
2477 * For each service unit a restart counter is now kept: it is increased
2478 each time the service is restarted due to Restart=, and may be
2479 queried using "systemctl show -p NRestarts …".
2481 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
2482 @signal and @timer have been added, for usage with SystemCallFilter=
2483 in unit files and the new --system-call-filter= command line option
2484 of systemd-nspawn (see above).
2486 * ExecStart= lines in unit files gained two new modifiers: when a
2487 command line is prefixed with "!" the command will be executed as
2488 configured, except for the credentials applied by
2489 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
2490 "+", but does still apply namespacing options unlike "+". There's
2491 also "!!" now, which is mostly identical, but becomes a NOP on
2492 systems that support ambient capabilities. This is useful to write
2493 unit files that work with ambient capabilities where possible but
2494 automatically fall back to traditional privilege dropping mechanisms
2495 on systems where this is not supported.
2497 * ListenNetlink= settings in socket units now support RDMA netlink
2500 * A new unit file setting LockPersonality= has been added which permits
2501 locking down the chosen execution domain ("personality") of a service
2504 * A new special target "getty-pre.target" has been added, which is
2505 ordered before all text logins, and may be used to order services
2506 before textual logins acquire access to the console.
2508 * systemd will now attempt to load the virtio-rng.ko kernel module very
2509 early on if a VM environment supporting this is detected. This should
2510 improve entropy during early boot in virtualized environments.
2512 * A _netdev option is now supported in /etc/crypttab that operates in a
2513 similar way as the same option in /etc/fstab: it permits configuring
2514 encrypted devices that need to be ordered after the network is up.
2515 Following this logic, two new special targets
2516 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
2517 added that are to cryptsetup.target what remote-fs.target and
2518 remote-fs-pre.target are to local-fs.target.
2520 * Service units gained a new UnsetEnvironment= setting which permits
2521 unsetting specific environment variables for services that are
2522 normally passed to it (for example in order to mask out locale
2523 settings for specific services that can't deal with it).
2525 * Units acquired a new boolean option IPAccounting=. When turned on, IP
2526 traffic accounting (packet count as well as byte count) is done for
2527 the service, and shown as part of "systemctl status" or "systemd-run
2530 * Service units acquired two new options IPAddressAllow= and
2531 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
2532 for configuring a simple IP access control list for all sockets of
2533 the unit. These options are available also on .slice and .socket
2534 units, permitting flexible access list configuration for individual
2535 services as well as groups of services (as defined by a slice unit),
2536 including system-wide. Note that IP ACLs configured this way are
2537 enforced on every single IPv4 and IPv6 socket created by any process
2538 of the service unit, and apply to ingress as well as egress traffic.
2540 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
2541 structured log message is generated each time the unit is stopped,
2542 containing information about the consumed resources of this
2545 * A new setting KeyringMode= has been added to unit files, which may be
2546 used to control how the kernel keyring is set up for executed
2549 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
2550 "systemctl kexec" and "systemctl exit" are now always asynchronous in
2551 behaviour (that is: these commands return immediately after the
2552 operation was enqueued instead of waiting for the operation to
2553 complete). Previously, "systemctl poweroff" and "systemctl reboot"
2554 were asynchronous on systems using systemd-logind (i.e. almost
2555 always, and like they were on sysvinit), and the other three commands
2556 were unconditionally synchronous. With this release this is cleaned
2557 up, and callers will see the same asynchronous behaviour on all
2558 systems for all five operations.
2560 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
2563 * .timer units now accept calendar specifications in other timezones
2564 than UTC or the local timezone.
2566 * The tmpfiles snippet var.conf has been changed to create
2567 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
2568 the "utmp" group already, and it appears to be generally understood
2569 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
2570 databases. Previously this was implemented correctly for all these
2571 databases excepts btmp, which has been opened up like this now
2572 too. Note that while the other databases are world-readable
2573 (i.e. 0644), btmp is not and remains more restrictive.
2575 * The systemd-resolve tool gained a new --reset-server-features
2576 switch. When invoked like this systemd-resolved will forget
2577 everything it learnt about the features supported by the configured
2578 upstream DNS servers, and restarts the feature probing logic on the
2579 next resolver look-up for them at the highest feature level
2582 * The status dump systemd-resolved sends to the logs upon receiving
2583 SIGUSR1 now also includes information about all DNS servers it is
2584 configured to use, and the features levels it probed for them.
2586 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
2587 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
2588 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
2589 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
2590 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
2591 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
2592 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
2593 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
2594 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
2595 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
2596 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
2597 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
2598 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
2599 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
2600 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
2601 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
2602 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
2603 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
2604 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
2605 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2607 — Berlin, 2017-10-06
2611 * Meson is now supported as build system in addition to Automake. It is
2612 our plan to remove Automake in one of our next releases, so that
2613 Meson becomes our exclusive build system. Hence, please start using
2614 the Meson build system in your downstream packaging. There's plenty
2615 of documentation around how to use Meson, the extremely brief
2618 ./autogen.sh && ./configure && make && sudo make install
2622 meson build && ninja -C build && sudo ninja -C build install
2624 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
2625 which permits configuring a timeout on the time a job is
2626 running. This is particularly useful for setting timeouts on jobs for
2629 * Unit files gained two new options ConditionUser= and ConditionGroup=
2630 for conditionalizing units based on the identity of the user/group
2631 running a systemd user instance.
2633 * systemd-networkd now understands a new FlowLabel= setting in the
2634 [VXLAN] section of .network files, as well as a Priority= in
2635 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
2636 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
2637 gained support for configuration of GENEVE links, and IPv6 address
2638 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
2640 * .link files now understand a new Port= setting.
2642 * systemd-networkd's DHCP support gained support for DHCP option 119
2643 (domain search list).
2645 * systemd-networkd gained support for serving IPv6 address ranges using
2646 the Router Advertisement protocol. The new .network configuration
2647 section [IPv6Prefix] may be used to configure the ranges to
2648 serve. This is implemented based on a new, minimal, native server
2649 implementation of RA.
2651 * journalctl's --output= switch gained support for a new parameter
2652 "short-iso-precise" for a mode where timestamps are shown as precise
2655 * systemd-udevd's "net_id" builtin may now generate stable network
2656 interface names from IBM PowerVM VIO devices as well as ACPI platform
2659 * MulticastDNS support in systemd-resolved may now be explicitly
2660 enabled/disabled using the new MulticastDNS= configuration file
2663 * systemd-resolved may now optionally use libidn2 instead of the libidn
2664 for processing internationalized domain names. Support for libidn2
2665 should be considered experimental and should not be enabled by
2668 * "machinectl pull-tar" and related call may now do verification of
2669 downloaded images using SUSE-style .sha256 checksum files in addition
2670 to the already existing support for validating using Ubuntu-style
2673 * sd-bus gained support for a new sd_bus_message_appendv() call which
2674 is va_list equivalent of sd_bus_message_append().
2676 * sd-boot gained support for validating images using SHIM/MOK.
2678 * The SMACK code learnt support for "onlycap".
2680 * systemd-mount --umount is now much smarter in figuring out how to
2681 properly unmount a device given its mount or device path.
2683 * The code to call libnss_dns as a fallback from libnss_resolve when
2684 the communication with systemd-resolved fails was removed. This
2685 fallback was redundant and interfered with the [!UNAVAIL=return]
2686 suffix. See nss-resolve(8) for the recommended configuration.
2688 * systemd-logind may now be restarted without losing state. It stores
2689 the file descriptors for devices it manages in the system manager
2690 using the FDSTORE= mechanism. Please note that further changes in
2691 other components may be required to make use of this (for example
2692 Xorg has code to listen for stops of systemd-logind and terminate
2693 itself when logind is stopped or restarted, in order to avoid using
2694 stale file descriptors for graphical devices, which is now
2695 counterproductive and must be reverted in order for restarts of
2696 systemd-logind to be safe. See
2697 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
2699 * All kernel-install plugins are called with the environment variable
2700 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
2701 /etc/machine-id. If the machine ID could not be determined,
2702 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
2703 anything in the entry directory (passed as the second argument) if
2704 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
2705 temporary directory is passed as the entry directory and removed
2706 after all the plugins exit.
2708 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
2709 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
2710 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
2711 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
2712 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
2713 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
2714 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
2715 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
2716 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
2717 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
2718 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
2719 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
2720 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
2721 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
2722 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
2723 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
2724 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
2725 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
2726 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
2727 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
2728 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
2729 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
2730 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
2731 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
2732 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
2733 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
2734 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
2735 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
2736 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
2739 — Berlin, 2017-07-12
2743 * The "hybrid" control group mode has been modified to improve
2744 compatibility with "legacy" cgroups-v1 setups. Specifically, the
2745 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
2746 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
2747 cgroups-v1 hierarchy), the only externally visible change being that
2748 the cgroups-v2 hierarchy is also mounted, to
2749 /sys/fs/cgroup/unified. This should provide a large degree of
2750 compatibility with "legacy" cgroups-v1, while taking benefit of the
2751 better management capabilities of cgroups-v2.
2753 * The default control group setup mode may be selected both a boot-time
2754 via a set of kernel command line parameters (specifically:
2755 systemd.unified_cgroup_hierarchy= and
2756 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
2757 default selected on the configure command line
2758 (--with-default-hierarchy=). The upstream default is "hybrid"
2759 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
2760 this will change in a future systemd version to be "unified" (pure
2761 cgroups-v2 mode). The third option for the compile time option is
2762 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
2763 distributions to default to "hybrid" mode for release distributions,
2764 starting with v233. We recommend "unified" for development
2765 distributions (specifically: distributions such as Fedora's rawhide)
2766 as that's where things are headed in the long run. Use "legacy" for
2767 greatest stability and compatibility only.
2769 * Note one current limitation of "unified" and "hybrid" control group
2770 setup modes: the kernel currently does not permit the systemd --user
2771 instance (i.e. unprivileged code) to migrate processes between two
2772 disconnected cgroup subtrees, even if both are managed and owned by
2773 the user. This effectively means "systemd-run --user --scope" doesn't
2774 work when invoked from outside of any "systemd --user" service or
2775 scope. Specifically, it is not supported from session scopes. We are
2776 working on fixing this in a future systemd version. (See #3388 for
2777 further details about this.)
2779 * DBus policy files are now installed into /usr rather than /etc. Make
2780 sure your system has dbus >= 1.9.18 running before upgrading to this
2781 version, or override the install path with --with-dbuspolicydir= .
2783 * All python scripts shipped with systemd (specifically: the various
2784 tests written in Python) now require Python 3.
2786 * systemd unit tests can now run standalone (without the source or
2787 build directories), and can be installed into /usr/lib/systemd/tests/
2788 with 'make install-tests'.
2790 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
2791 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
2794 * Support for the %c, %r, %R specifiers in unit files has been
2795 removed. Specifiers are not supposed to be dependent on configuration
2796 in the unit file itself (so that they resolve the same regardless
2797 where used in the unit files), but these specifiers were influenced
2798 by the Slice= option.
2800 * The shell invoked by debug-shell.service now defaults to /bin/sh in
2801 all cases. If distributions want to use a different shell for this
2802 purpose (for example Fedora's /sbin/sushell) they need to specify
2803 this explicitly at configure time using --with-debug-shell=.
2805 * The confirmation spawn prompt has been reworked to offer the
2808 (c)ontinue, proceed without asking anymore
2809 (D)ump, show the state of the unit
2810 (f)ail, don't execute the command and pretend it failed
2812 (i)nfo, show a short summary of the unit
2813 (j)obs, show jobs that are in progress
2814 (s)kip, don't execute the command and pretend it succeeded
2815 (y)es, execute the command
2817 The 'n' choice for the confirmation spawn prompt has been removed,
2818 because its meaning was confusing.
2820 The prompt may now also be redirected to an alternative console by
2821 specifying the console as parameter to systemd.confirm_spawn=.
2823 * Services of Type=notify require a READY=1 notification to be sent
2824 during startup. If no such message is sent, the service now fails,
2825 even if the main process exited with a successful exit code.
2827 * Services that fail to start up correctly now always have their
2828 ExecStopPost= commands executed. Previously, they'd enter "failed"
2829 state directly, without executing these commands.
2831 * The option MulticastDNS= of network configuration files has acquired
2832 an actual implementation. With MulticastDNS=yes a host can resolve
2833 names of remote hosts and reply to mDNS A and AAAA requests.
2835 * When units are about to be started an additional check is now done to
2836 ensure that all dependencies of type BindsTo= (when used in
2837 combination with After=) have been started.
2839 * systemd-analyze gained a new verb "syscall-filter" which shows which
2840 system call groups are defined for the SystemCallFilter= unit file
2841 setting, and which system calls they contain.
2843 * A new system call filter group "@filesystem" has been added,
2844 consisting of various file system related system calls. Group
2845 "@reboot" has been added, covering reboot, kexec and shutdown related
2846 calls. Finally, group "@swap" has been added covering swap
2847 configuration related calls.
2849 * A new unit file option RestrictNamespaces= has been added that may be
2850 used to restrict access to the various process namespace types the
2851 Linux kernel provides. Specifically, it may be used to take away the
2852 right for a service unit to create additional file system, network,
2853 user, and other namespaces. This sandboxing option is particularly
2854 relevant due to the high amount of recently discovered namespacing
2855 related vulnerabilities in the kernel.
2857 * systemd-udev's .link files gained support for a new AutoNegotiation=
2858 setting for configuring Ethernet auto-negotiation.
2860 * systemd-networkd's .network files gained support for a new
2861 ListenPort= setting in the [DHCP] section to explicitly configure the
2862 UDP client port the DHCP client shall listen on.
2864 * .network files gained a new Unmanaged= boolean setting for explicitly
2865 excluding one or more interfaces from management by systemd-networkd.
2867 * The systemd-networkd ProxyARP= option has been renamed to
2868 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
2869 renamed to ReduceARPProxy=. The old names continue to be available
2872 * systemd-networkd gained support for configuring IPv6 Proxy NDP
2873 addresses via the new IPv6ProxyNDPAddress= .network file setting.
2875 * systemd-networkd's bonding device support gained support for two new
2876 configuration options ActiveSlave= and PrimarySlave=.
2878 * The various options in the [Match] section of .network files gained
2879 support for negative matching.
2881 * New systemd-specific mount options are now understood in /etc/fstab:
2883 x-systemd.mount-timeout= may be used to configure the maximum
2884 permitted runtime of the mount command.
2886 x-systemd.device-bound may be set to bind a mount point to its
2887 backing device unit, in order to automatically remove a mount point
2888 if its backing device is unplugged. This option may also be
2889 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
2890 on the block device, which is now automatically set for all CDROM
2891 drives, so that mounted CDs are automatically unmounted when they are
2892 removed from the drive.
2894 x-systemd.after= and x-systemd.before= may be used to explicitly
2895 order a mount after or before another unit or mount point.
2897 * Enqueued start jobs for device units are now automatically garbage
2898 collected if there are no jobs waiting for them anymore.
2900 * systemctl list-jobs gained two new switches: with --after, for every
2901 queued job the jobs it's waiting for are shown; with --before the
2902 jobs which it's blocking are shown.
2904 * systemd-nspawn gained support for ephemeral boots from disk images
2905 (or in other words: --ephemeral and --image= may now be
2906 combined). Moreover, ephemeral boots are now supported for normal
2907 directories, even if the backing file system is not btrfs. Of course,
2908 if the file system does not support file system snapshots or
2909 reflinks, the initial copy operation will be relatively expensive, but
2910 this should still be suitable for many use cases.
2912 * Calendar time specifications in .timer units now support
2913 specifications relative to the end of a month by using "~" instead of
2914 "-" as separator between month and day. For example, "*-02~03" means
2915 "the third last day in February". In addition a new syntax for
2916 repeated events has been added using the "/" character. For example,
2917 "9..17/2:00" means "every two hours from 9am to 5pm".
2919 * systemd-socket-proxyd gained a new parameter --connections-max= for
2920 configuring the maximum number of concurrent connections.
2922 * sd-id128 gained a new API for generating unique IDs for the host in a
2923 way that does not leak the machine ID. Specifically,
2924 sd_id128_get_machine_app_specific() derives an ID based on the
2925 machine ID a in well-defined, non-reversible, stable way. This is
2926 useful whenever an identifier for the host is needed but where the
2927 identifier shall not be useful to identify the system beyond the
2928 scope of the application itself. (Internally this uses HMAC-SHA256 as
2929 keyed hash function using the machine ID as input.)
2931 * NotifyAccess= gained a new supported value "exec". When set
2932 notifications are accepted from all processes systemd itself invoked,
2933 including all control processes.
2935 * .nspawn files gained support for defining overlay mounts using the
2936 Overlay= and OverlayReadOnly= options. Previously this functionality
2937 was only available on the systemd-nspawn command line.
2939 * systemd-nspawn's --bind= and --overlay= options gained support for
2940 bind/overlay mounts whose source lies within the container tree by
2941 prefixing the source path with "+".
2943 * systemd-nspawn's --bind= and --overlay= options gained support for
2944 automatically allocating a temporary source directory in /var/tmp
2945 that is removed when the container dies. Specifically, if the source
2946 directory is specified as empty string this mechanism is selected. An
2947 example usage is --overlay=+/var::/var, which creates an overlay
2948 mount based on the original /var contained in the image, overlayed
2949 with a temporary directory in the host's /var/tmp. This way changes
2950 to /var are automatically flushed when the container shuts down.
2952 * systemd-nspawn --image= option does now permit raw file system block
2953 devices (in addition to images containing partition tables, as
2956 * The disk image dissection logic in systemd-nspawn gained support for
2957 automatically setting up LUKS encrypted as well as Verity protected
2958 partitions. When a container is booted from an encrypted image the
2959 passphrase is queried at start-up time. When a container with Verity
2960 data is started, the root hash is search in a ".roothash" file
2961 accompanying the disk image (alternatively, pass the root hash via
2962 the new --root-hash= command line option).
2964 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
2965 be used to dissect disk images the same way as systemd-nspawn does
2966 it, following the Bootable Partition Specification. It may even be
2967 used to mount disk images with complex partition setups (including
2968 LUKS and Verity partitions) to a local host directory, in order to
2969 inspect them. This tool is not considered public API (yet), and is
2970 thus not installed into /usr/bin. Please do not rely on its
2971 existence, since it might go away or be changed in later systemd
2974 * A new generator "systemd-verity-generator" has been added, similar in
2975 style to "systemd-cryptsetup-generator", permitting automatic setup of
2976 Verity root partitions when systemd boots up. In order to make use of
2977 this your partition setup should follow the Discoverable Partitions
2978 Specification, and the GPT partition ID of the root file system
2979 partition should be identical to the upper 128bit of the Verity root
2980 hash. The GPT partition ID of the Verity partition protecting it
2981 should be the lower 128bit of the Verity root hash. If the partition
2982 image follows this model it is sufficient to specify a single
2983 "roothash=" kernel command line argument to both configure which root
2984 image and verity partition to use as well as the root hash for
2985 it. Note that systemd-nspawn's Verity support follows the same
2986 semantics, meaning that disk images with proper Verity data in place
2987 may be booted in containers with systemd-nspawn as well as on
2988 physical systems via the verity generator. Also note that the "mkosi"
2989 tool available at https://github.com/systemd/mkosi has been updated
2990 to generate Verity protected disk images following this scheme. In
2991 fact, it has been updated to generate disk images that optionally
2992 implement a complete UEFI SecureBoot trust chain, involving a signed
2993 kernel and initrd image that incorporates such a root hash as well as
2994 a Verity-enabled root partition.
2996 * The hardware database (hwdb) udev supports has been updated to carry
2997 accelerometer quirks.
2999 * All system services are now run with a fresh kernel keyring set up
3000 for them. The invocation ID is stored by default in it, thus
3001 providing a safe, non-overridable way to determine the invocation
3004 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
3005 options for bind mounting arbitrary paths in a service-specific
3006 way. When these options are used, arbitrary host or service files and
3007 directories may be mounted to arbitrary locations in the service's
3010 * Documentation has been added that lists all of systemd's low-level
3011 environment variables:
3013 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
3015 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
3016 whether a specific socket file descriptor matches a specified socket
3019 * systemd-firstboot has been updated to check for the
3020 systemd.firstboot= kernel command line option. It accepts a boolean
3021 and when set to false the first boot questions are skipped.
3023 * systemd-fstab-generator has been updated to check for the
3024 systemd.volatile= kernel command line option, which either takes an
3025 optional boolean parameter or the special value "state". If used the
3026 system may be booted in a "volatile" boot mode. Specifically,
3027 "systemd.volatile" is used, the root directory will be mounted as
3028 tmpfs, and only /usr is mounted from the actual root file system. If
3029 "systemd.volatile=state" is used, the root directory will be mounted
3030 as usual, but /var is mounted as tmpfs. This concept provides similar
3031 functionality as systemd-nspawn's --volatile= option, but provides it
3032 on physical boots. Use this option for implementing stateless
3033 systems, or testing systems with all state and/or configuration reset
3034 to the defaults. (Note though that many distributions are not
3035 prepared to boot up without a populated /etc or /var, though.)
3037 * systemd-gpt-auto-generator gained support for LUKS encrypted root
3038 partitions. Previously it only supported LUKS encrypted partitions
3039 for all other uses, except for the root partition itself.
3041 * Socket units gained support for listening on AF_VSOCK sockets for
3042 communication in virtualized QEMU environments.
3044 * The "configure" script gained a new option --with-fallback-hostname=
3045 for specifying the fallback hostname to use if none is configured in
3046 /etc/hostname. For example, by specifying
3047 --with-fallback-hostname=fedora it is possible to default to a
3048 hostname of "fedora" on pristine installations.
3050 * systemd-cgls gained support for a new --unit= switch for listing only
3051 the control groups of a specific unit. Similar --user-unit= has been
3052 added for listing only the control groups of a specific user unit.
3054 * systemd-mount gained a new --umount switch for unmounting a mount or
3055 automount point (and all mount/automount points below it).
3057 * systemd will now refuse full configuration reloads (via systemctl
3058 daemon-reload and related calls) unless at least 16MiB of free space
3059 are available in /run. This is a safety precaution in order to ensure
3060 that generators can safely operate after the reload completed.
3062 * A new unit file option RootImage= has been added, which has a similar
3063 effect as RootDirectory= but mounts the service's root directory from
3064 a disk image instead of plain directory. This logic reuses the same
3065 image dissection and mount logic that systemd-nspawn already uses,
3066 and hence supports any disk images systemd-nspawn supports, including
3067 those following the Discoverable Partition Specification, as well as
3068 Verity enabled images. This option enables systemd to run system
3069 services directly off disk images acting as resource bundles,
3070 possibly even including full integrity data.
3072 * A new MountAPIVFS= unit file option has been added, taking a boolean
3073 argument. If enabled /proc, /sys and /dev (collectively called the
3074 "API VFS") will be mounted for the service. This is only relevant if
3075 RootDirectory= or RootImage= is used for the service, as these mounts
3076 are of course in place in the host mount namespace anyway.
3078 * systemd-nspawn gained support for a new --pivot-root= switch. If
3079 specified the root directory within the container image is pivoted to
3080 the specified mount point, while the original root disk is moved to a
3081 different place. This option enables booting of ostree images
3082 directly with systemd-nspawn.
3084 * The systemd build scripts will no longer complain if the NTP server
3085 addresses are not changed from the defaults. Google now supports
3086 these NTP servers officially. We still recommend downstreams to
3087 properly register an NTP pool with the NTP pool project though.
3089 * coredumpctl gained a new "--reverse" option for printing the list
3090 of coredumps in reverse order.
3092 * coredumpctl will now show additional information about truncated and
3093 inaccessible coredumps, as well as coredumps that are still being
3094 processed. It also gained a new --quiet switch for suppressing
3095 additional informational message in its output.
3097 * coredumpctl gained support for only showing coredumps newer and/or
3098 older than specific timestamps, using the new --since= and --until=
3099 options, reminiscent of journalctl's options by the same name.
3101 * The systemd-coredump logic has been improved so that it may be reused
3102 to collect backtraces in non-compiled languages, for example in
3103 scripting languages such as Python.
3105 * machinectl will now show the UID shift of local containers, if user
3106 namespacing is enabled for them.
3108 * systemd will now optionally run "environment generator" binaries at
3109 configuration load time. They may be used to add environment
3110 variables to the environment block passed to services invoked. One
3111 user environment generator is shipped by default that sets up
3112 environment variables based on files dropped into /etc/environment.d
3113 and ~/.config/environment.d/.
3115 * systemd-resolved now includes the new, recently published 2017 DNSSEC
3118 * hostnamed has been updated to report a new chassis type of
3119 "convertible" to cover "foldable" laptops that can both act as a
3120 tablet and as a laptop, such as various Lenovo Yoga devices.
3122 Contributions from: Adrián López, Alexander Galanin, Alexander
3123 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
3124 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
3125 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
3126 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
3127 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
3128 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
3129 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
3130 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
3131 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
3132 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
3133 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
3134 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
3135 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
3136 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
3137 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
3138 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
3139 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
3140 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
3141 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
3142 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
3143 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
3144 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
3145 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
3146 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
3147 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
3148 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
3149 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
3152 — Berlin, 2017-03-01
3156 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
3157 RestrictAddressFamilies= enabled. These sandboxing options should
3158 generally be compatible with the various external udev call-out
3159 binaries we are aware of, however there may be exceptions, in
3160 particular when exotic languages for these call-outs are used. In
3161 this case, consider turning off these settings locally.
3163 * The new RemoveIPC= option can be used to remove IPC objects owned by
3164 the user or group of a service when that service exits.
3166 * The new ProtectKernelModules= option can be used to disable explicit
3167 load and unload operations of kernel modules by a service. In
3168 addition access to /usr/lib/modules is removed if this option is set.
3170 * ProtectSystem= option gained a new value "strict", which causes the
3171 whole file system tree with the exception of /dev, /proc, and /sys,
3172 to be remounted read-only for a service.
3174 * The new ProtectKernelTunables= option can be used to disable
3175 modification of configuration files in /sys and /proc by a service.
3176 Various directories and files are remounted read-only, so access is
3177 restricted even if the file permissions would allow it.
3179 * The new ProtectControlGroups= option can be used to disable write
3180 access by a service to /sys/fs/cgroup.
3182 * Various systemd services have been hardened with
3183 ProtectKernelTunables=yes, ProtectControlGroups=yes,
3184 RestrictAddressFamilies=.
3186 * Support for dynamically creating users for the lifetime of a service
3187 has been added. If DynamicUser=yes is specified, user and group IDs
3188 will be allocated from the range 61184..65519 for the lifetime of the
3189 service. They can be resolved using the new nss-systemd.so NSS
3190 module. The module must be enabled in /etc/nsswitch.conf. Services
3191 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
3192 any resources allocated by the service will be cleaned up when the
3193 service exits. They also have ProtectHome=read-only and
3194 ProtectSystem=strict enabled, so they are not able to make any
3195 permanent modifications to the system.
3197 * The nss-systemd module also always resolves root and nobody, making
3198 it possible to have no /etc/passwd or /etc/group files in minimal
3199 container or chroot environments.
3201 * Services may be started with their own user namespace using the new
3202 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
3203 under which the service is running are mapped. All other users are
3206 * Support for the cgroup namespace has been added to systemd-nspawn. If
3207 supported by kernel, the container system started by systemd-nspawn
3208 will have its own view of the cgroup hierarchy. This new behaviour
3209 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
3211 * The new MemorySwapMax= option can be used to limit the maximum swap
3212 usage under the unified cgroup hierarchy.
3214 * Support for the CPU controller in the unified cgroup hierarchy has
3215 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
3216 options. This controller requires out-of-tree patches for the kernel
3217 and the support is provisional.
3219 * Mount and automount units may now be created transiently
3220 (i.e. dynamically at runtime via the bus API, instead of requiring
3221 unit files in the file system).
3223 * systemd-mount is a new tool which may mount file systems – much like
3224 mount(8), optionally pulling in additional dependencies through
3225 transient .mount and .automount units. For example, this tool
3226 automatically runs fsck on a backing block device before mounting,
3227 and allows the automount logic to be used dynamically from the
3228 command line for establishing mount points. This tool is particularly
3229 useful when dealing with removable media, as it will ensure fsck is
3230 run – if necessary – before the first access and that the file system
3231 is quickly unmounted after each access by utilizing the automount
3232 logic. This maximizes the chance that the file system on the
3233 removable media stays in a clean state, and if it isn't in a clean
3234 state is fixed automatically.
3236 * LazyUnmount=yes option for mount units has been added to expose the
3237 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
3240 * /efi will be used as the mount point of the EFI boot partition, if
3241 the directory is present, and the mount point was not configured
3242 through other means (e.g. fstab). If /efi directory does not exist,
3243 /boot will be used as before. This makes it easier to automatically
3244 mount the EFI partition on systems where /boot is used for something
3247 * When operating on GPT disk images for containers, systemd-nspawn will
3248 now mount the ESP to /boot or /efi according to the same rules as PID
3249 1 running on a host. This allows tools like "bootctl" to operate
3250 correctly within such containers, in order to make container images
3251 bootable on physical systems.
3253 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
3255 * Two new user session targets have been added to support running
3256 graphical sessions under the systemd --user instance:
3257 graphical-session.target and graphical-session-pre.target. See
3258 systemd.special(7) for a description of how those targets should be
3261 * The vconsole initialization code has been significantly reworked to
3262 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
3263 support unicode keymaps. Font and keymap configuration will now be
3264 copied to all allocated virtual consoles.
3266 * FreeBSD's bhyve virtualization is now detected.
3268 * Information recorded in the journal for core dumps now includes the
3269 contents of /proc/mountinfo and the command line of the process at
3270 the top of the process hierarchy (which is usually the init process
3273 * systemd-journal-gatewayd learned the --directory= option to serve
3274 files from the specified location.
3276 * journalctl --root=… can be used to peruse the journal in the
3277 /var/log/ directories inside of a container tree. This is similar to
3278 the existing --machine= option, but does not require the container to
3281 * The hardware database has been extended to support
3282 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
3285 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
3286 specify the click rate for mice which include a horizontal wheel with
3287 a click rate that is different than the one for the vertical wheel.
3289 * systemd-run gained a new --wait option that makes service execution
3290 synchronous. (Specifically, the command will not return until the
3291 specified service binary exited.)
3293 * systemctl gained a new --wait option that causes the start command to
3294 wait until the units being started have terminated again.
3296 * A new journal output mode "short-full" has been added which displays
3297 timestamps with abbreviated English day names and adds a timezone
3298 suffix. Those timestamps include more information than the default
3299 "short" output mode, and can be passed directly to journalctl's
3300 --since= and --until= options.
3302 * /etc/resolv.conf will be bind-mounted into containers started by
3303 systemd-nspawn, if possible, so any changes to resolv.conf contents
3304 are automatically propagated to the container.
3306 * The number of instances for socket-activated services originating
3307 from a single IP address can be limited with
3308 MaxConnectionsPerSource=, extending the existing setting of
3311 * systemd-networkd gained support for vcan ("Virtual CAN") interface
3314 * .netdev and .network configuration can now be extended through
3317 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
3318 Segmentation Offload, Generic Receive Offload, Large Receive Offload
3319 can be enabled and disabled using the new UDPSegmentationOffload=,
3320 TCPSegmentationOffload=, GenericSegmentationOffload=,
3321 GenericReceiveOffload=, LargeReceiveOffload= options in the
3322 [Link] section of .link files.
3324 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
3325 Port VLAN ID can be configured for bridge devices using the new STP=,
3326 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
3327 section of .netdev files.
3329 * The route table to which routes received over DHCP or RA should be
3330 added can be configured with the new RouteTable= option in the [DHCP]
3331 and [IPv6AcceptRA] sections of .network files.
3333 * The Address Resolution Protocol can be disabled on links managed by
3334 systemd-networkd using the ARP=no setting in the [Link] section of
3337 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
3338 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
3339 encode information about the result and exit codes of the current
3340 service runtime cycle.
3342 * systemd-sysctl will now configure kernel parameters in the order
3343 they occur in the configuration files. This matches what sysctl
3344 has been traditionally doing.
3346 * kernel-install "plugins" that are executed to perform various
3347 tasks after a new kernel is added and before an old one is removed
3348 can now return a special value to terminate the procedure and
3349 prevent any later plugins from running.
3351 * Journald's SplitMode=login setting has been deprecated. It has been
3352 removed from documentation, and its use is discouraged. In a future
3353 release it will be completely removed, and made equivalent to current
3354 default of SplitMode=uid.
3356 * Storage=both option setting in /etc/systemd/coredump.conf has been
3357 removed. With fast LZ4 compression storing the core dump twice is not
3360 * The --share-system systemd-nspawn option has been replaced with an
3361 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
3362 this functionality is discouraged. In addition the variables
3363 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
3364 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
3365 individual namespaces.
3367 * "machinectl list" now shows the IP address of running containers in
3368 the output, as well as OS release information.
3370 * "loginctl list" now shows the TTY of each session in the output.
3372 * sd-bus gained new API calls sd_bus_track_set_recursive(),
3373 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
3374 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
3375 tracking objects in a "recursive" mode, where a single client can be
3376 counted multiple times, if it takes multiple references.
3378 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
3379 sd_bus_get_exit_on_disconnect(). They may be used to make a
3380 process using sd-bus automatically exit if the bus connection is
3383 * Bus clients of the service manager may now "pin" loaded units into
3384 memory, by taking an explicit reference on them. This is useful to
3385 ensure the client can retrieve runtime data about the service even
3386 after the service completed execution. Taking such a reference is
3387 available only for privileged clients and should be helpful to watch
3388 running services in a race-free manner, and in particular collect
3389 information about exit statuses and results.
3391 * The nss-resolve module has been changed to strictly return UNAVAIL
3392 when communication via D-Bus with resolved failed, and NOTFOUND when
3393 a lookup completed but was negative. This means it is now possible to
3394 neatly configure fallbacks using nsswitch.conf result checking
3395 expressions. Taking benefit of this, the new recommended
3396 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
3398 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
3400 * A new setting CtrlAltDelBurstAction= has been added to
3401 /etc/systemd/system.conf which may be used to configure the precise
3402 behaviour if the user on the console presses Ctrl-Alt-Del more often
3403 than 7 times in 2s. Previously this would unconditionally result in
3404 an expedited, immediate reboot. With this new setting the precise
3405 operation may be configured in more detail, and also turned off
3408 * In .netdev files two new settings RemoteChecksumTx= and
3409 RemoteChecksumRx= are now understood that permit configuring the
3410 remote checksumming logic for VXLAN networks.
3412 * The service manager learnt a new "invocation ID" concept for invoked
3413 services. Each runtime cycle of a service will get a new invocation
3414 ID (a 128bit random UUID) assigned that identifies the current
3415 run of the service uniquely and globally. A new invocation ID
3416 is generated each time a service starts up. The journal will store
3417 the invocation ID of a service along with any logged messages, thus
3418 making the invocation ID useful for matching the online runtime of a
3419 service with the offline log data it generated in a safe way without
3420 relying on synchronized timestamps. In many ways this new service
3421 invocation ID concept is similar to the kernel's boot ID concept that
3422 uniquely and globally identifies the runtime of each boot. The
3423 invocation ID of a service is passed to the service itself via an
3424 environment variable ($INVOCATION_ID). A new bus call
3425 GetUnitByInvocationID() has been added that is similar to GetUnit()
3426 but instead of retrieving the bus path for a unit by its name
3427 retrieves it by its invocation ID. The returned path is valid only as
3428 long as the passed invocation ID is current.
3430 * systemd-resolved gained a new "DNSStubListener" setting in
3431 resolved.conf. It either takes a boolean value or the special values
3432 "udp" and "tcp", and configures whether to enable the stub DNS
3433 listener on 127.0.0.53:53.
3435 * IP addresses configured via networkd may now carry additional
3436 configuration settings supported by the kernel. New options include:
3437 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
3438 PrefixRoute=, AutoJoin=.
3440 * The PAM configuration fragment file for "user@.service" shipped with
3441 systemd (i.e. the --user instance of systemd) has been stripped to
3442 the minimum necessary to make the system boot. Previously, it
3443 contained Fedora-specific stanzas that did not apply to other
3444 distributions. It is expected that downstream distributions add
3445 additional configuration lines, matching their needs to this file,
3446 using it only as rough template of what systemd itself needs. Note
3447 that this reduced fragment does not even include an invocation of
3448 pam_limits which most distributions probably want to add, even though
3449 systemd itself does not need it. (There's also the new build time
3450 option --with-pamconfdir=no to disable installation of the PAM
3453 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
3454 capability is now also dropped from its set (in addition to
3455 CAP_SYS_MKNOD as before).
3457 * In service unit files it is now possible to connect a specific named
3458 file descriptor with stdin/stdout/stdout of an executed service. The
3459 name may be specified in matching .socket units using the
3460 FileDescriptorName= setting.
3462 * A number of journal settings may now be configured on the kernel
3463 command line. Specifically, the following options are now understood:
3464 systemd.journald.max_level_console=,
3465 systemd.journald.max_level_store=,
3466 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
3467 systemd.journald.max_level_wall=.
3469 * "systemctl is-enabled --full" will now show by which symlinks a unit
3470 file is enabled in the unit dependency tree.
3472 * Support for VeraCrypt encrypted partitions has been added to the
3473 "cryptsetup" logic and /etc/crypttab.
3475 * systemd-detect-virt gained support for a new --private-users switch
3476 that checks whether the invoking processes are running inside a user
3477 namespace. Similar, a new special value "private-users" for the
3478 existing ConditionVirtualization= setting has been added, permitting
3479 skipping of specific units in user namespace environments.
3481 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
3482 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
3483 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
3484 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
3485 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
3486 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
3487 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
3488 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
3489 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
3490 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
3491 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
3492 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
3493 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
3494 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
3495 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
3496 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
3497 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
3498 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
3499 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
3500 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
3501 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
3502 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
3503 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
3504 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
3505 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
3506 Jędrzejewski-Szmek, Zeal Jagannatha
3508 — Santa Fe, 2016-11-03
3512 * In service units the various ExecXYZ= settings have been extended
3513 with an additional special character as first argument of the
3514 assigned value: if the character '+' is used the specified command
3515 line it will be run with full privileges, regardless of User=,
3516 Group=, CapabilityBoundingSet= and similar options. The effect is
3517 similar to the existing PermissionsStartOnly= option, but allows
3518 configuration of this concept for each executed command line
3521 * Services may now alter the service watchdog timeout at runtime by
3522 sending a WATCHDOG_USEC= message via sd_notify().
3524 * MemoryLimit= and related unit settings now optionally take percentage
3525 specifications. The percentage is taken relative to the amount of
3526 physical memory in the system (or in case of containers, the assigned
3527 amount of memory). This allows scaling service resources neatly with
3528 the amount of RAM available on the system. Similarly, systemd-logind's
3529 RuntimeDirectorySize= option now also optionally takes percentage
3532 * In similar fashion TasksMax= takes percentage values now, too. The
3533 value is taken relative to the configured maximum number of processes
3534 on the system. The per-service task maximum has been changed to 15%
3535 using this functionality. (Effectively this is an increase of 512 →
3536 4915 for service units, given the kernel's default pid_max setting.)
3538 * Calendar time specifications in .timer units now understand a ".."
3539 syntax for time ranges. Example: "4..7:10" may now be used for
3540 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
3543 * The InaccessableDirectories=, ReadOnlyDirectories= and
3544 ReadWriteDirectories= unit file settings have been renamed to
3545 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
3546 applied to all kinds of file nodes, and not just directories, with
3547 the exception of symlinks. Specifically these settings may now be
3548 used on block and character device nodes, UNIX sockets and FIFOS as
3549 well as regular files. The old names of these settings remain
3550 available for compatibility.
3552 * systemd will now log about all service processes it kills forcibly
3553 (using SIGKILL) because they remained after the clean shutdown phase
3554 of the service completed. This should help identifying services that
3555 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
3556 systemd-logind's configuration a similar log message is generated for
3557 processes killed at the end of each session due to this setting.
3559 * systemd will now set the $JOURNAL_STREAM environment variable for all
3560 services whose stdout/stderr are connected to the Journal (which
3561 effectively means by default: all services). The variable contains
3562 the device and inode number of the file descriptor used for
3563 stdout/stderr. This may be used by invoked programs to detect whether
3564 their stdout/stderr is connected to the Journal, in which case they
3565 can switch over to direct Journal communication, thus being able to
3566 pass extended, structured metadata along with their log messages. As
3567 one example, this is now used by glib's logging primitives.
3569 * When using systemd's default tmp.mount unit for /tmp, the mount point
3570 will now be established with the "nosuid" and "nodev" options. This
3571 avoids privilege escalation attacks that put traps and exploits into
3572 /tmp. However, this might cause problems if you e. g. put container
3573 images or overlays into /tmp; if you need this, override tmp.mount's
3574 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
3577 * systemd now supports the "memory" cgroup controller also on
3580 * The systemd-cgtop tool now optionally takes a control group path as
3581 command line argument. If specified, the control group list shown is
3582 limited to subgroups of that group.
3584 * The SystemCallFilter= unit file setting gained support for
3585 pre-defined, named system call filter sets. For example
3586 SystemCallFilter=@clock is now an effective way to make all clock
3587 changing-related system calls unavailable to a service. A number of
3588 similar pre-defined groups are defined. Writing system call filters
3589 for system services is simplified substantially with this new
3590 concept. Accordingly, all of systemd's own, long-running services now
3591 enable system call filtering based on this, by default.
3593 * A new service setting MemoryDenyWriteExecute= has been added, taking
3594 a boolean value. If turned on, a service may no longer create memory
3595 mappings that are writable and executable at the same time. This
3596 enhances security for services where this is enabled as it becomes
3597 harder to dynamically write and then execute memory in exploited
3598 service processes. This option has been enabled for all of systemd's
3599 own long-running services.
3601 * A new RestrictRealtime= service setting has been added, taking a
3602 boolean argument. If set the service's processes may no longer
3603 acquire realtime scheduling. This improves security as realtime
3604 scheduling may otherwise be used to easily freeze the system.
3606 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
3607 value. This may be used for requesting that the system manager inside
3608 of the container reports start-up completion to nspawn which then
3609 propagates this notification further to the service manager
3610 supervising nspawn itself. A related option NotifyReady= in .nspawn
3611 files has been added too. This functionality allows ordering of the
3612 start-up of multiple containers using the usual systemd ordering
3615 * machinectl gained a new command "stop" that is an alias for
3618 * systemd-resolved gained support for contacting DNS servers on
3619 link-local IPv6 addresses.
3621 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
3622 its caches. A method call for requesting the same operation has been
3623 added to the bus API too, and is made available via "systemd-resolve
3626 * systemd-resolve gained a new --status switch. If passed a brief
3627 summary of the used DNS configuration with per-interface information
3630 * resolved.conf gained a new Cache= boolean option, defaulting to
3631 on. If turned off local DNS caching is disabled. This comes with a
3632 performance penalty in particular when DNSSEC is enabled. Note that
3633 resolved disables its internal caching implicitly anyway, when the
3634 configured DNS server is on a host-local IP address such as ::1 or
3635 127.0.0.1, thus automatically avoiding double local caching.
3637 * systemd-resolved now listens on the local IP address 127.0.0.53:53
3638 for DNS requests. This improves compatibility with local programs
3639 that do not use the libc NSS or systemd-resolved's bus APIs for name
3640 resolution. This minimal DNS service is only available to local
3641 programs and does not implement the full DNS protocol, but enough to
3642 cover local DNS clients. A new, static resolv.conf file, listing just
3643 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
3644 now recommended to make /etc/resolv.conf a symlink to this file in
3645 order to route all DNS lookups to systemd-resolved, regardless if
3646 done via NSS, the bus API or raw DNS packets. Note that this local
3647 DNS service is not as fully featured as the libc NSS or
3648 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
3649 used to deliver link-local address information (as this implies
3650 sending a local interface index along), LLMNR/mDNS support via this
3651 interface is severely restricted. It is thus strongly recommended for
3652 all applications to use the libc NSS API or native systemd-resolved
3655 * systemd-networkd's bridge support learned a new setting
3656 VLANFiltering= for controlling VLAN filtering. Moreover a new section
3657 in .network files has been added for configuring VLAN bridging in
3658 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
3660 * systemd-networkd's IPv6 Router Advertisement code now makes use of
3661 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
3662 now be acquired without relying on DHCPv6. Two new options
3663 UseDomains= and UseDNS= have been added to configure this behaviour.
3665 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
3666 renamed IPv6AcceptRA=, without altering its behaviour. The old
3667 setting name remains available for compatibility reasons.
3669 * The systemd-networkd VTI/VTI6 tunneling support gained new options
3670 Key=, InputKey= and OutputKey=.
3672 * systemd-networkd gained support for VRF ("Virtual Routing Function")
3673 interface configuration.
3675 * "systemctl edit" may now be used to create new unit files by
3676 specifying the --force switch.
3678 * sd-event gained a new function sd_event_get_iteration() for
3679 requesting the current iteration counter of the event loop. It starts
3680 at zero and is increased by one with each event loop iteration.
3682 * A new rpm macro %systemd_ordering is provided by the macros.systemd
3683 file. It can be used in lieu of %systemd_requires in packages which
3684 don't use any systemd functionality and are intended to be installed
3685 in minimal containers without systemd present. This macro provides
3686 ordering dependencies to ensure that if the package is installed in
3687 the same rpm transaction as systemd, systemd will be installed before
3688 the scriptlets for the package are executed, allowing unit presets
3691 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
3692 been added to simplify packaging of generators.
3694 * The os-release file gained VERSION_CODENAME field for the
3695 distribution nickname (e.g. VERSION_CODENAME=woody).
3697 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
3698 can be set to disable parsing of metadata and the creation
3699 of persistent symlinks for that device.
3701 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
3702 to make them available to logged-in users has been reverted.
3704 * Much of the common code of the various systemd components is now
3705 built into an internal shared library libsystemd-shared-231.so
3706 (incorporating the systemd version number in the name, to be updated
3707 with future releases) that the components link to. This should
3708 decrease systemd footprint both in memory during runtime and on
3709 disk. Note that the shared library is not for public use, and is
3710 neither API nor ABI stable, but is likely to change with every new
3711 released update. Packagers need to make sure that binaries
3712 linking to libsystemd-shared.so are updated in step with the
3715 * Configuration for "mkosi" is now part of the systemd
3716 repository. mkosi is a tool to easily build legacy-free OS images,
3717 and is available on github: https://github.com/systemd/mkosi. If
3718 "mkosi" is invoked in the build tree a new raw OS image is generated
3719 incorporating the systemd sources currently being worked on and a
3720 clean, fresh distribution installation. The generated OS image may be
3721 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
3722 UEFI PC. This functionality is particularly useful to easily test
3723 local changes made to systemd in a pristine, defined environment. See
3724 doc/HACKING for details.
3726 * configure learned the --with-support-url= option to specify the
3727 distribution's bugtracker.
3729 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
3730 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
3731 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
3732 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
3733 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
3734 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
3735 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
3736 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
3737 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
3738 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
3739 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
3740 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
3741 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
3742 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
3743 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
3744 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
3745 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
3746 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
3747 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
3749 — Berlin, 2016-07-25
3753 * DNSSEC is now turned on by default in systemd-resolved (in
3754 "allow-downgrade" mode), but may be turned off during compile time by
3755 passing "--with-default-dnssec=no" to "configure" (and of course,
3756 during runtime with DNSSEC= in resolved.conf). We recommend
3757 downstreams to leave this on at least during development cycles and
3758 report any issues with the DNSSEC logic upstream. We are very
3759 interested in collecting feedback about the DNSSEC validator and its
3760 limitations in the wild. Note however, that DNSSEC support is
3761 probably nothing downstreams should turn on in stable distros just
3762 yet, as it might create incompatibilities with a few DNS servers and
3763 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
3764 automatically whenever we detect such incompatible setups, but there
3765 might be systems we do not cover yet. Hence: please help us testing
3766 the DNSSEC code, leave this on where you can, report back, but then
3767 again don't consider turning this on in your stable, LTS or
3768 production release just yet. (Note that you have to enable
3769 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
3770 and its DNSSEC mode for host name resolution from local
3773 * systemd-resolve conveniently resolves DANE records with the --tlsa
3774 option and OPENPGPKEY records with the --openpgp option. It also
3775 supports dumping raw DNS record data via the new --raw= switch.
3777 * systemd-logind will now by default terminate user processes that are
3778 part of the user session scope unit (session-XX.scope) when the user
3779 logs out. This behavior is controlled by the KillUserProcesses=
3780 setting in logind.conf, and the previous default of "no" is now
3781 changed to "yes". This means that user sessions will be properly
3782 cleaned up after, but additional steps are necessary to allow
3783 intentionally long-running processes to survive logout.
3785 While the user is logged in at least once, user@.service is running,
3786 and any service that should survive the end of any individual login
3787 session can be started at a user service or scope using systemd-run.
3788 systemd-run(1) man page has been extended with an example which shows
3789 how to run screen in a scope unit underneath user@.service. The same
3790 command works for tmux.
3792 After the user logs out of all sessions, user@.service will be
3793 terminated too, by default, unless the user has "lingering" enabled.
3794 To effectively allow users to run long-term tasks even if they are
3795 logged out, lingering must be enabled for them. See loginctl(1) for
3796 details. The default polkit policy was modified to allow users to
3797 set lingering for themselves without authentication.
3799 Previous defaults can be restored at compile time by the
3800 --without-kill-user-processes option to "configure".
3802 * systemd-logind gained new configuration settings SessionsMax= and
3803 InhibitorsMax=, both with a default of 8192. It will not register new
3804 user sessions or inhibitors above this limit.
3806 * systemd-logind will now reload configuration on SIGHUP.
3808 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
3809 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
3810 enable. Also, support for the "io" cgroup controller in the unified
3811 hierarchy has been added, so that the "memory", "pids" and "io" are
3812 now the controllers that are supported on the unified hierarchy.
3814 WARNING: it is not possible to use previous systemd versions with
3815 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
3816 is necessary to also update systemd in the initramfs if using the
3817 unified hierarchy. An updated SELinux policy is also required.
3819 * LLDP support has been extended, and both passive (receive-only) and
3820 active (sender) modes are supported. Passive mode ("routers-only") is
3821 enabled by default in systemd-networkd. Active LLDP mode is enabled
3822 by default for containers on the internal network. The "networkctl
3823 lldp" command may be used to list information gathered. "networkctl
3824 status" will also show basic LLDP information on connected peers now.
3826 * The IAID and DUID unique identifier sent in DHCP requests may now be
3827 configured for the system and each .network file managed by
3828 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
3830 * systemd-networkd gained support for configuring proxy ARP support for
3831 each interface, via the ProxyArp= setting in .network files. It also
3832 gained support for configuring the multicast querier feature of
3833 bridge devices, via the new MulticastQuerier= setting in .netdev
3834 files. Similarly, snooping on the IGMP traffic can be controlled
3835 via the new setting MulticastSnooping=.
3837 A new setting PreferredLifetime= has been added for addresses
3838 configured in .network file to configure the lifetime intended for an
3841 The systemd-networkd DHCP server gained the option EmitRouter=, which
3842 defaults to yes, to configure whether the DHCP Option 3 (Router)
3845 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
3846 systemd-socket-activate and installed into /usr/bin. It is now fully
3849 * systemd-journald now uses separate threads to flush changes to disk
3850 when closing journal files, thus reducing impact of slow disk I/O on
3851 logging performance.
3853 * The sd-journal API gained two new calls
3854 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
3855 can be used to open journal files using file descriptors instead of
3856 file or directory paths. sd_journal_open_container() has been
3857 deprecated, sd_journal_open_directory_fd() should be used instead
3858 with the flag SD_JOURNAL_OS_ROOT.
3860 * journalctl learned a new output mode "-o short-unix" that outputs log
3861 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
3862 UTC). It also gained support for a new --no-hostname setting to
3863 suppress the hostname column in the family of "short" output modes.
3865 * systemd-ask-password now optionally skips printing of the password to
3866 stdout with --no-output which can be useful in scripts.
3868 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
3869 (devices tagged with ID_MAKER_TOOL) are now tagged with
3870 "uaccess" and are available to logged in users.
3872 * The DeviceAllow= unit setting now supports specifiers (with "%").
3874 * "systemctl show" gained a new --value switch, which allows print a
3875 only the contents of a specific unit property, without also printing
3876 the property's name. Similar support was added to "show*" verbs
3877 of loginctl and machinectl that output "key=value" lists.
3879 * A new unit type "generated" was added for files dynamically generated
3880 by generator tools. Similarly, a new unit type "transient" is used
3881 for unit files created using the runtime API. "systemctl enable" will
3882 refuse to operate on such files.
3884 * A new command "systemctl revert" has been added that may be used to
3885 revert to the vendor version of a unit file, in case local changes
3886 have been made by adding drop-ins or overriding the unit file.
3888 * "machinectl clean" gained a new verb to automatically remove all or
3889 just hidden container images.
3891 * systemd-tmpfiles gained support for a new line type "e" for emptying
3892 directories, if they exist, without creating them if they don't.
3894 * systemd-nspawn gained support for automatically patching the UID/GIDs
3895 of the owners and the ACLs of all files and directories in a
3896 container tree to match the UID/GID user namespacing range selected
3897 for the container invocation. This mode is enabled via the new
3898 --private-users-chown switch. It also gained support for
3899 automatically choosing a free, previously unused UID/GID range when
3900 starting a container, via the new --private-users=pick setting (which
3901 implies --private-users-chown). Together, these options for the first
3902 time make user namespacing for nspawn containers fully automatic and
3903 thus deployable. The systemd-nspawn@.service template unit file has
3904 been changed to use this functionality by default.
3906 * systemd-nspawn gained a new --network-zone= switch, that allows
3907 creating ad-hoc virtual Ethernet links between multiple containers,
3908 that only exist as long as at least one container referencing them is
3909 running. This allows easy connecting of multiple containers with a
3910 common link that implements an Ethernet broadcast domain. Each of
3911 these network "zones" may be named relatively freely by the user, and
3912 may be referenced by any number of containers, but each container may
3913 only reference one of these "zones". On the lower level, this is
3914 implemented by an automatically managed bridge network interface for
3915 each zone, that is created when the first container referencing its
3916 zone is created and removed when the last one referencing its zone
3919 * The default start timeout may now be configured on the kernel command
3920 line via systemd.default_timeout_start_sec=. It was already
3921 configurable via the DefaultTimeoutStartSec= option in
3922 /etc/systemd/system.conf.
3924 * Socket units gained a new TriggerLimitIntervalSec= and
3925 TriggerLimitBurst= setting to configure a limit on the activation
3926 rate of the socket unit.
3928 * The LimitNICE= setting now optionally takes normal UNIX nice values
3929 in addition to the raw integer limit value. If the specified
3930 parameter is prefixed with "+" or "-" and is in the range -20..19 the
3931 value is understood as UNIX nice value. If not prefixed like this it
3932 is understood as raw RLIMIT_NICE limit.
3934 * Note that the effect of the PrivateDevices= unit file setting changed
3935 slightly with this release: the per-device /dev file system will be
3936 mounted read-only from this version on, and will have "noexec"
3937 set. This (minor) change of behavior might cause some (exceptional)
3938 legacy software to break, when PrivateDevices=yes is set for its
3939 service. Please leave PrivateDevices= off if you run into problems
3942 * systemd-bootchart has been split out to a separate repository:
3943 https://github.com/systemd/systemd-bootchart
3945 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
3946 merged into the kernel in its current form.
3948 * The compatibility libraries libsystemd-daemon.so,
3949 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
3950 which have been deprecated since systemd-209 have been removed along
3951 with the corresponding pkg-config files. All symbols provided by
3952 those libraries are provided by libsystemd.so.
3954 * The Capabilities= unit file setting has been removed (it is ignored
3955 for backwards compatibility). AmbientCapabilities= and
3956 CapabilityBoundingSet= should be used instead.
3958 * A new special target has been added, initrd-root-device.target,
3959 which creates a synchronization point for dependencies of the root
3960 device in early userspace. Initramfs builders must ensure that this
3961 target is now included in early userspace.
3963 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
3964 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
3965 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
3966 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
3967 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
3968 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
3969 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
3970 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
3971 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
3972 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
3973 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
3974 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
3975 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
3976 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
3977 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
3978 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
3979 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
3980 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
3981 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
3982 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
3983 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
3984 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
3985 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
3986 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
3987 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
3990 — Fairfax, 2016-05-21
3994 * The systemd-resolved DNS resolver service has gained a substantial
3995 set of new features, most prominently it may now act as a DNSSEC
3996 validating stub resolver. DNSSEC mode is currently turned off by
3997 default, but is expected to be turned on by default in one of the
3998 next releases. For now, we invite everybody to test the DNSSEC logic
3999 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
4000 service also gained a full set of D-Bus interfaces, including calls
4001 to configure DNS and DNSSEC settings per link (for use by external
4002 network management software). systemd-resolved and systemd-networkd
4003 now distinguish between "search" and "routing" domains. The former
4004 are used to qualify single-label names, the latter are used purely
4005 for routing lookups within certain domains to specific links.
4006 resolved now also synthesizes RRs for all entries from /etc/hosts.
4008 * The systemd-resolve tool (which is a client utility for
4009 systemd-resolved) has been improved considerably and is now fully
4010 supported and documented. Hence it has moved from /usr/lib/systemd to
4013 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
4016 * The coredump collection logic has been reworked: when a coredump is
4017 collected it is now written to disk, compressed and processed
4018 (including stacktrace extraction) from a new instantiated service
4019 systemd-coredump@.service, instead of directly from the
4020 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
4021 processing large coredumps can take up a substantial amount of
4022 resources and time, and this previously happened entirely outside of
4023 systemd's service supervision. With the new logic the core_pattern
4024 hook only does minimal metadata collection before passing off control
4025 to the new instantiated service, which is configured with a time
4026 limit, a nice level and other settings to minimize negative impact on
4027 the rest of the system. Also note that the new logic will honour the
4028 RLIMIT_CORE setting of the crashed process, which now allows users
4029 and processes to turn off coredumping for their processes by setting
4032 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
4033 and all forked processes by default. Previously, PID 1 would leave
4034 the setting at "0" for all processes, as set by the kernel. Note that
4035 the resource limit traditionally has no effect on the generated
4036 coredumps on the system if the /proc/sys/kernel/core_pattern hook
4037 logic is used. Since the limit is now honoured (see above) its
4038 default has been changed so that the coredumping logic is enabled by
4039 default for all processes, while allowing specific opt-out.
4041 * When the stacktrace is extracted from processes of system users, this
4042 is now done as "systemd-coredump" user, in order to sandbox this
4043 potentially security sensitive parsing operation. (Note that when
4044 processing coredumps of normal users this is done under the user ID
4045 of process that crashed, as before.) Packagers should take notice
4046 that it is now necessary to create the "systemd-coredump" system user
4047 and group at package installation time.
4049 * The systemd-activate socket activation testing tool gained support
4050 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
4051 and --seqpacket switches. It also has been extended to support both
4052 new-style and inetd-style file descriptor passing. Use the new
4053 --inetd switch to request inetd-style file descriptor passing.
4055 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
4056 variable, which takes a boolean value. If set to false, ANSI color
4057 output is disabled in the tools even when run on a terminal that
4060 * The VXLAN support in networkd now supports two new settings
4061 DestinationPort= and PortRange=.
4063 * A new systemd.machine_id= kernel command line switch has been added,
4064 that may be used to set the machine ID in /etc/machine-id if it is
4065 not initialized yet. This command line option has no effect if the
4066 file is already initialized.
4068 * systemd-nspawn gained a new --as-pid2 switch that invokes any
4069 specified command line as PID 2 rather than PID 1 in the
4070 container. In this mode PID 1 is a minimal stub init process that
4071 implements the special POSIX and Linux semantics of PID 1 regarding
4072 signal and child process management. Note that this stub init process
4073 is implemented in nspawn itself and requires no support from the
4074 container image. This new logic is useful to support running
4075 arbitrary commands in the container, as normal processes are
4076 generally not prepared to run as PID 1.
4078 * systemd-nspawn gained a new --chdir= switch for setting the current
4079 working directory for the process started in the container.
4081 * "journalctl /dev/sda" will now output all kernel log messages for
4082 specified device from the current boot, in addition to all devices
4083 that are parents of it. This should make log output about devices
4084 pretty useful, as long as kernel drivers attach enough metadata to
4085 the log messages. (The usual SATA drivers do.)
4087 * The sd-journal API gained two new calls
4088 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
4089 that report whether log data from /run or /var has been found.
4091 * journalctl gained a new switch "--fields" that prints all journal
4092 record field names currently in use in the journal. This is backed
4093 by two new sd-journal API calls sd_journal_enumerate_fields() and
4094 sd_journal_restart_fields().
4096 * Most configurable timeouts in systemd now expect an argument of
4097 "infinity" to turn them off, instead of "0" as before. The semantics
4098 from now on is that a timeout of "0" means "now", and "infinity"
4099 means "never". To maintain backwards compatibility, "0" continues to
4100 turn off previously existing timeout settings.
4102 * "systemctl reload-or-try-restart" has been renamed to "systemctl
4103 try-reload-or-restart" to clarify what it actually does: the "try"
4104 logic applies to both reloading and restarting, not just restarting.
4105 The old name continues to be accepted for compatibility.
4107 * On boot-up, when PID 1 detects that the system clock is behind the
4108 release date of the systemd version in use, the clock is now set
4109 to the latter. Previously, this was already done in timesyncd, in order
4110 to avoid running with clocks set to the various clock epochs such as
4111 1902, 1938 or 1970. With this change the logic is now done in PID 1
4112 in addition to timesyncd during early boot-up, so that it is enforced
4113 before the first process is spawned by systemd. Note that the logic
4114 in timesyncd remains, as it is more comprehensive and ensures
4115 clock monotonicity by maintaining a persistent timestamp file in
4116 /var. Since /var is generally not available in earliest boot or the
4117 initrd, this part of the logic remains in timesyncd, and is not done
4120 * Support for tweaking details in net_cls.class_id through the
4121 NetClass= configuration directive has been removed, as the kernel
4122 people have decided to deprecate that controller in cgroup v2.
4123 Userspace tools such as nftables are moving over to setting rules
4124 that are specific to the full cgroup path of a task, which obsoletes
4125 these controllers anyway. The NetClass= directive is kept around for
4126 legacy compatibility reasons. For a more in-depth description of the
4127 kernel change, please refer to the respective upstream commit:
4129 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
4131 * A new service setting RuntimeMaxSec= has been added that may be used
4132 to specify a maximum runtime for a service. If the timeout is hit, the
4133 service is terminated and put into a failure state.
4135 * A new service setting AmbientCapabilities= has been added. It allows
4136 configuration of additional Linux process capabilities that are
4137 passed to the activated processes. This is only available on very
4140 * The process resource limit settings in service units may now be used
4141 to configure hard and soft limits individually.
4143 * The various libsystemd APIs such as sd-bus or sd-event now publicly
4144 expose support for gcc's __attribute__((cleanup())) C extension.
4145 Specifically, for many object destructor functions alternative
4146 versions have been added that have names suffixed with "p" and take a
4147 pointer to a pointer to the object to destroy, instead of just a
4148 pointer to the object itself. This is useful because these destructor
4149 functions may be used directly as parameters to the cleanup
4150 construct. Internally, systemd has been a heavy user of this GCC
4151 extension for a long time, and with this change similar support is
4152 now available to consumers of the library outside of systemd. Note
4153 that by using this extension in your sources compatibility with old
4154 and strictly ANSI compatible C compilers is lost. However, all gcc or
4155 LLVM versions of recent years support this extension.
4157 * Timer units gained support for a new setting RandomizedDelaySec= that
4158 allows configuring some additional randomized delay to the configured
4159 time. This is useful to spread out timer events to avoid load peaks in
4160 clusters or larger setups.
4162 * Calendar time specifications now support sub-second accuracy.
4164 * Socket units now support listening on SCTP and UDP-lite protocol
4167 * The sd-event API now comes with a full set of man pages.
4169 * Older versions of systemd contained experimental support for
4170 compressing journal files and coredumps with the LZ4 compressor that
4171 was not compatible with the lz4 binary (due to API limitations of the
4172 lz4 library). This support has been removed; only support for files
4173 compatible with the lz4 binary remains. This LZ4 logic is now
4174 officially supported and no longer considered experimental.
4176 * The dkr image import logic has been removed again from importd. dkr's
4177 micro-services focus doesn't fit into the machine image focus of
4178 importd, and quickly got out of date with the upstream dkr API.
4180 * Creation of the /run/lock/lockdev/ directory was dropped from
4181 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
4182 been available for many years. If you still need this, you need to
4183 create your own tmpfiles.d config file with:
4185 d /run/lock/lockdev 0775 root lock -
4187 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
4188 and RebootArgument= have been moved from the [Service] section of
4189 unit files to [Unit], and they are now supported on all unit types,
4190 not just service units. Of course, systemd will continue to
4191 understand these settings also at the old location, in order to
4192 maintain compatibility.
4194 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
4195 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
4196 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
4197 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
4198 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
4199 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
4200 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
4201 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
4202 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
4203 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
4204 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
4205 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
4206 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
4207 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
4208 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
4209 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
4210 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
4211 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
4212 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
4214 — Berlin, 2016-02-11
4218 * A number of properties previously only settable in unit
4219 files are now also available as properties to set when
4220 creating transient units programmatically via the bus, as it
4221 is exposed with systemd-run's --property=
4222 setting. Specifically, these are: SyslogIdentifier=,
4223 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
4224 EnvironmentFile=, ReadWriteDirectories=,
4225 ReadOnlyDirectories=, InaccessibleDirectories=,
4226 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
4228 * When creating transient services via the bus API it is now
4229 possible to pass in a set of file descriptors to use as
4230 STDIN/STDOUT/STDERR for the invoked process.
4232 * Slice units may now be created transiently via the bus APIs,
4233 similar to the way service and scope units may already be
4234 created transiently.
4236 * Wherever systemd expects a calendar timestamp specification
4237 (like in journalctl's --since= and --until= switches) UTC
4238 timestamps are now supported. Timestamps suffixed with "UTC"
4239 are now considered to be in Universal Time Coordinated
4240 instead of the local timezone. Also, timestamps may now
4241 optionally be specified with sub-second accuracy. Both of
4242 these additions also apply to recurring calendar event
4243 specification, such as OnCalendar= in timer units.
4245 * journalctl gained a new "--sync" switch that asks the
4246 journal daemon to write all so far unwritten log messages to
4247 disk and sync the files, before returning.
4249 * systemd-tmpfiles learned two new line types "q" and "Q" that
4250 operate like "v", but also set up a basic btrfs quota
4251 hierarchy when used on a btrfs file system with quota
4254 * tmpfiles' "v", "q" and "Q" will now create a plain directory
4255 instead of a subvolume (even on a btrfs file system) if the
4256 root directory is a plain directory, and not a
4257 subvolume. This should simplify things with certain chroot()
4258 environments which are not aware of the concept of btrfs
4261 * systemd-detect-virt gained a new --chroot switch to detect
4262 whether execution takes place in a chroot() environment.
4264 * CPUAffinity= now takes CPU index ranges in addition to
4267 * The various memory-related resource limit settings (such as
4268 LimitAS=) now understand the usual K, M, G, ... suffixes to
4269 the base of 1024 (IEC). Similar, the time-related resource
4270 limit settings understand the usual min, h, day, ...
4273 * There's a new system.conf setting DefaultTasksMax= to
4274 control the default TasksMax= setting for services and
4275 scopes running on the system. (TasksMax= is the primary
4276 setting that exposes the "pids" cgroup controller on systemd
4277 and was introduced in the previous systemd release.) The
4278 setting now defaults to 512, which means services that are
4279 not explicitly configured otherwise will only be able to
4280 create 512 processes or threads at maximum, from this
4281 version on. Note that this means that thread- or
4282 process-heavy services might need to be reconfigured to set
4283 TasksMax= to a higher value. It is sufficient to set
4284 TasksMax= in these specific unit files to a higher value, or
4285 even "infinity". Similar, there's now a logind.conf setting
4286 UserTasksMax= that defaults to 4096 and limits the total
4287 number of processes or tasks each user may own
4288 concurrently. nspawn containers also have the TasksMax=
4289 value set by default now, to 8192. Note that all of this
4290 only has an effect if the "pids" cgroup controller is
4291 enabled in the kernel. The general benefit of these changes
4292 should be a more robust and safer system, that provides a
4293 certain amount of per-service fork() bomb protection.
4295 * systemd-nspawn gained the new --network-veth-extra= switch
4296 to define additional and arbitrarily-named virtual Ethernet
4297 links between the host and the container.
4299 * A new service execution setting PassEnvironment= has been
4300 added that allows importing select environment variables
4301 from PID1's environment block into the environment block of
4304 * Timer units gained support for a new RemainAfterElapse=
4305 setting which takes a boolean argument. It defaults to on,
4306 exposing behaviour unchanged to previous releases. If set to
4307 off, timer units are unloaded after they elapsed if they
4308 cannot elapse again. This is particularly useful for
4309 transient timer units, which shall not stay around longer
4310 than until they first elapse.
4312 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
4313 default now (the kernel default is 16). This is beneficial
4314 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
4315 allows substantially larger numbers of queued
4316 datagrams. This should increase the capability of systemd to
4317 parallelize boot-up, as logging and sd_notify() are unlikely
4318 to stall execution anymore. If you need to change the value
4319 from the new defaults, use the usual sysctl.d/ snippets.
4321 * The compression framing format used by the journal or
4322 coredump processing has changed to be in line with what the
4323 official LZ4 tools generate. LZ4 compression support in
4324 systemd was considered unsupported previously, as the format
4325 was not compatible with the normal tools. With this release
4326 this has changed now, and it is hence safe for downstream
4327 distributions to turn it on. While not compressing as well
4328 as the XZ, LZ4 is substantially faster, which makes
4329 it a good default choice for the compression logic in the
4330 journal and in coredump handling.
4332 * Any reference to /etc/mtab has been dropped from
4333 systemd. The file has been obsolete since a while, but
4334 systemd refused to work on systems where it was incorrectly
4335 set up (it should be a symlink or non-existent). Please make
4336 sure to update to util-linux 2.27.1 or newer in conjunction
4337 with this systemd release, which also drops any reference to
4338 /etc/mtab. If you maintain a distribution make sure that no
4339 software you package still references it, as this is a
4340 likely source of bugs. There's also a glibc bug pending,
4341 asking for removal of any reference to this obsolete file:
4343 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
4345 Note that only util-linux versions built with
4346 --enable-libmount-force-mountinfo are supported.
4348 * Support for the ".snapshot" unit type has been removed. This
4349 feature turned out to be little useful and little used, and
4350 has now been removed from the core and from systemctl.
4352 * The dependency types RequiresOverridable= and
4353 RequisiteOverridable= have been removed from systemd. They
4354 have been used only very sparingly to our knowledge and
4355 other options that provide a similar effect (such as
4356 systemctl --mode=ignore-dependencies) are much more useful
4357 and commonly used. Moreover, they were only half-way
4358 implemented as the option to control behaviour regarding
4359 these dependencies was never added to systemctl. By removing
4360 these dependency types the execution engine becomes a bit
4361 simpler. Unit files that use these dependencies should be
4362 changed to use the non-Overridable dependency types
4363 instead. In fact, when parsing unit files with these
4364 options, that's what systemd will automatically convert them
4365 too, but it will also warn, asking users to fix the unit
4366 files accordingly. Removal of these dependency types should
4367 only affect a negligible number of unit files in the wild.
4369 * Behaviour of networkd's IPForward= option changed
4370 (again). It will no longer maintain a per-interface setting,
4371 but propagate one way from interfaces where this is enabled
4372 to the global kernel setting. The global setting will be
4373 enabled when requested by a network that is set up, but
4374 never be disabled again. This change was made to make sure
4375 IPv4 and IPv6 behaviour regarding packet forwarding is
4376 similar (as the Linux IPv6 stack does not support
4377 per-interface control of this setting) and to minimize
4380 * In unit files the behaviour of %u, %U, %h, %s has
4381 changed. These specifiers will now unconditionally resolve
4382 to the various user database fields of the user that the
4383 systemd instance is running as, instead of the user
4384 configured in the specific unit via User=. Note that this
4385 effectively doesn't change much, as resolving of these
4386 specifiers was already turned off in the --system instance
4387 of systemd, as we cannot do NSS lookups from PID 1. In the
4388 --user instance of systemd these specifiers where correctly
4389 resolved, but hardly made any sense, since the user instance
4390 lacks privileges to do user switches anyway, and User= is
4391 hence useless. Moreover, even in the --user instance of
4392 systemd behaviour was awkward as it would only take settings
4393 from User= assignment placed before the specifier into
4394 account. In order to unify and simplify the logic around
4395 this the specifiers will now always resolve to the
4396 credentials of the user invoking the manager (which in case
4397 of PID 1 is the root user).
4399 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
4400 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
4401 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
4402 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
4403 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
4404 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
4405 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
4406 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
4407 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
4408 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
4409 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
4410 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
4411 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
4412 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
4415 — Berlin, 2015-11-18
4419 * systemd now depends on util-linux v2.27. More specifically,
4420 the newly added mount monitor feature in libmount now
4421 replaces systemd's former own implementation.
4423 * libmount mandates /etc/mtab not to be regular file, and
4424 systemd now enforces this condition at early boot.
4425 /etc/mtab has been deprecated and warned about for a very
4426 long time, so systems running systemd should already have
4427 stopped having this file around as anything else than a
4428 symlink to /proc/self/mounts.
4430 * Support for the "pids" cgroup controller has been added. It
4431 allows accounting the number of tasks in a cgroup and
4432 enforcing limits on it. This adds two new setting
4433 TasksAccounting= and TasksMax= to each unit, as well as a
4434 global option DefaultTasksAccounting=.
4436 * Support for the "net_cls" cgroup controller has been added.
4437 It allows assigning a net class ID to each task in the
4438 cgroup, which can then be used in firewall rules and traffic
4439 shaping configurations. Note that the kernel netfilter net
4440 class code does not currently work reliably for ingress
4441 packets on unestablished sockets.
4443 This adds a new config directive called NetClass= to CGroup
4444 enabled units. Allowed values are positive numbers for fixed
4445 assignments and "auto" for picking a free value
4448 * 'systemctl is-system-running' now returns 'offline' if the
4449 system is not booted with systemd. This command can now be
4450 used as a substitute for 'systemd-notify --booted'.
4452 * Watchdog timeouts have been increased to 3 minutes for all
4453 in-tree service files. Apparently, disk IO issues are more
4454 frequent than we hoped, and user reported >1 minute waiting
4457 * 'machine-id-commit' functionality has been merged into
4458 'machine-id-setup --commit'. The separate binary has been
4461 * The WorkingDirectory= directive in unit files may now be set
4462 to the special value '~'. In this case, the working
4463 directory is set to the home directory of the user
4464 configured in User=.
4466 * "machinectl shell" will now open the shell in the home
4467 directory of the selected user by default.
4469 * The CrashChVT= configuration file setting is renamed to
4470 CrashChangeVT=, following our usual logic of not
4471 abbreviating unnecessarily. The old directive is still
4472 supported for compat reasons. Also, this directive now takes
4473 an integer value between 1 and 63, or a boolean value. The
4474 formerly supported '-1' value for disabling stays around for
4477 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
4478 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
4479 RootDirectory= properties can now be set for transient
4482 * The systemd-analyze tool gained a new "set-log-target" verb
4483 to change the logging target the system manager logs to
4484 dynamically during runtime. This is similar to how
4485 "systemd-analyze set-log-level" already changes the log
4488 * In nspawn /sys is now mounted as tmpfs, with only a selected
4489 set of subdirectories mounted in from the real sysfs. This
4490 enhances security slightly, and is useful for ensuring user
4491 namespaces work correctly.
4493 * Support for USB FunctionFS activation has been added. This
4494 allows implementation of USB gadget services that are
4495 activated as soon as they are requested, so that they don't
4496 have to run continuously, similar to classic socket
4499 * The "systemctl exit" command now optionally takes an
4500 additional parameter that sets the exit code to return from
4501 the systemd manager when exiting. This is only relevant when
4502 running the systemd user instance, or when running the
4503 system instance in a container.
4505 * sd-bus gained the new API calls sd_bus_path_encode_many()
4506 and sd_bus_path_decode_many() that allow easy encoding and
4507 decoding of multiple identifier strings inside a D-Bus
4508 object path. Another new call sd_bus_default_flush_close()
4509 has been added to flush and close per-thread default
4512 * systemd-cgtop gained support for a -M/--machine= switch to
4513 show the control groups within a certain container only.
4515 * "systemctl kill" gained support for an optional --fail
4516 switch. If specified the requested operation will fail of no
4517 processes have been killed, because the unit had no
4518 processes attached, or similar.
4520 * A new systemd.crash_reboot=1 kernel command line option has
4521 been added that triggers a reboot after crashing. This can
4522 also be set through CrashReboot= in systemd.conf.
4524 * The RuntimeDirectory= setting now understands unit
4525 specifiers like %i or %f.
4527 * A new (still internal) library API sd-ipv4acd has been added,
4528 that implements address conflict detection for IPv4. It's
4529 based on code from sd-ipv4ll, and will be useful for
4530 detecting DHCP address conflicts.
4532 * File descriptors passed during socket activation may now be
4533 named. A new API sd_listen_fds_with_names() is added to
4534 access the names. The default names may be overridden,
4535 either in the .socket file using the FileDescriptorName=
4536 parameter, or by passing FDNAME= when storing the file
4537 descriptors using sd_notify().
4539 * systemd-networkd gained support for:
4541 - Setting the IPv6 Router Advertisement settings via
4542 IPv6AcceptRouterAdvertisements= in .network files.
4544 - Configuring the HelloTimeSec=, MaxAgeSec= and
4545 ForwardDelaySec= bridge parameters in .netdev files.
4547 - Configuring PreferredSource= for static routes in
4550 * The "ask-password" framework used to query for LUKS harddisk
4551 passwords or SSL passwords during boot gained support for
4552 caching passwords in the kernel keyring, if it is
4553 available. This makes sure that the user only has to type in
4554 a passphrase once if there are multiple objects to unlock
4555 with the same one. Previously, such password caching was
4556 available only when Plymouth was used; this moves the
4557 caching logic into the systemd codebase itself. The
4558 "systemd-ask-password" utility gained a new --keyname=
4559 switch to control which kernel keyring key to use for
4560 caching a password in. This functionality is also useful for
4561 enabling display managers such as gdm to automatically
4562 unlock the user's GNOME keyring if its passphrase, the
4563 user's password and the harddisk password are the same, if
4564 gdm-autologin is used.
4566 * When downloading tar or raw images using "machinectl
4567 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
4568 file is now also downloaded, if it is available and stored
4569 next to the image file.
4571 * Units of type ".socket" gained a new boolean setting
4572 Writable= which is only useful in conjunction with
4573 ListenSpecial=. If true, enables opening the specified
4574 special file in O_RDWR mode rather than O_RDONLY mode.
4576 * systemd-rfkill has been reworked to become a singleton
4577 service that is activated through /dev/rfkill on each rfkill
4578 state change and saves the settings to disk. This way,
4579 systemd-rfkill is now compatible with devices that exist
4580 only intermittendly, and even restores state if the previous
4581 system shutdown was abrupt rather than clean.
4583 * The journal daemon gained support for vacuuming old journal
4584 files controlled by the number of files that shall remain,
4585 in addition to the already existing control by size and by
4586 date. This is useful as journal interleaving performance
4587 degrades with too many separate journal files, and allows
4588 putting an effective limit on them. The new setting defaults
4589 to 100, but this may be changed by setting SystemMaxFiles=
4590 and RuntimeMaxFiles= in journald.conf. Also, the
4591 "journalctl" tool gained the new --vacuum-files= switch to
4592 manually vacuum journal files to leave only the specified
4593 number of files in place.
4595 * udev will now create /dev/disk/by-path links for ATA devices
4596 on kernels where that is supported.
4598 * Galician, Serbian, Turkish and Korean translations were added.
4600 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
4601 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
4602 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
4603 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
4604 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
4605 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
4606 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
4607 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
4608 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
4609 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
4610 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
4611 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
4612 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
4613 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
4614 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
4615 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
4616 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
4617 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
4619 — Berlin, 2015-10-07
4623 * The DHCP implementation of systemd-networkd gained a set of
4626 - The DHCP server now supports emitting DNS and NTP
4627 information. It may be enabled and configured via
4628 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
4629 and NTP information is enabled, but no servers are
4630 configured, the corresponding uplink information (if there
4631 is any) is propagated.
4633 - Server and client now support transmission and reception
4634 of timezone information. It can be configured via the
4635 newly introduced network options UseTimezone=,
4636 EmitTimezone=, and Timezone=. Transmission of timezone
4637 information is enabled between host and containers by
4638 default now: the container will change its local timezone
4639 to what the host has set.
4641 - Lease timeouts can now be configured via
4642 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
4644 - The DHCP server improved on the stability of
4645 leases. Clients are more likely to get the same lease
4646 information back, even if the server loses state.
4648 - The DHCP server supports two new configuration options to
4649 control the lease address pool metrics, PoolOffset= and
4652 * The encapsulation limit of tunnels in systemd-networkd may
4653 now be configured via 'EncapsulationLimit='. It allows
4654 modifying the maximum additional levels of encapsulation
4655 that are permitted to be prepended to a packet.
4657 * systemd now supports the concept of user buses replacing
4658 session buses, if used with dbus-1.10 (and enabled via dbus
4659 --enable-user-session). It previously only supported this on
4660 kdbus-enabled systems, and this release expands this to
4661 'dbus-daemon' systems.
4663 * systemd-networkd now supports predictable interface names
4666 * systemd now optionally supports the new Linux kernel
4667 "unified" control group hierarchy. If enabled via the kernel
4668 command-line option 'systemd.unified_cgroup_hierarchy=1',
4669 systemd will try to mount the unified cgroup hierarchy
4670 directly on /sys/fs/cgroup. If not enabled, or not
4671 available, systemd will fall back to the legacy cgroup
4672 hierarchy setup, as before. Host system and containers can
4673 mix and match legacy and unified hierarchies as they
4674 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
4675 environment variable to individually select the hierarchy to
4676 use for executed containers. By default, nspawn will use the
4677 unified hierarchy for the containers if the host uses the
4678 unified hierarchy, and the legacy hierarchy otherwise.
4679 Please note that at this point the unified hierarchy is an
4680 experimental kernel feature and is likely to change in one
4681 of the next kernel releases. Therefore, it should not be
4682 enabled by default in downstream distributions yet. The
4683 minimum required kernel version for the unified hierarchy to
4684 work is 4.2. Note that when the unified hierarchy is used
4685 for the first time delegated access to controllers is
4686 safe. Because of this systemd-nspawn containers will get
4687 access to controllers now, as will systemd user
4688 sessions. This means containers and user sessions may now
4689 manage their own resources, partitioning up what the system
4692 * A new special scope unit "init.scope" has been introduced
4693 that encapsulates PID 1 of the system. It may be used to
4694 determine resource usage and enforce resource limits on PID
4695 1 itself. PID 1 hence moved out of the root of the control
4698 * The cgtop tool gained support for filtering out kernel
4699 threads when counting tasks in a control group. Also, the
4700 count of processes is now recursively summed up by
4701 default. Two options -k and --recursive= have been added to
4702 revert to old behaviour. The tool has also been updated to
4703 work correctly in containers now.
4705 * systemd-nspawn's --bind= and --bind-ro= options have been
4706 extended to allow creation of non-recursive bind mounts.
4708 * libsystemd gained two new calls sd_pid_get_cgroup() and
4709 sd_peer_get_cgroup() which return the control group path of
4710 a process or peer of a connected AF_UNIX socket. This
4711 function call is particularly useful when implementing
4712 delegated subtrees support in the control group hierarchy.
4714 * The "sd-event" event loop API of libsystemd now supports
4715 correct dequeuing of real-time signals, without losing
4718 * When systemd requests a polkit decision when managing units it
4719 will now add additional fields to the request, including unit
4720 name and desired operation. This enables more powerful polkit
4721 policies, that make decisions depending on these parameters.
4723 * nspawn learnt support for .nspawn settings files, that may
4724 accompany the image files or directories of containers, and
4725 may contain additional settings for the container. This is
4726 an alternative to configuring container parameters via the
4727 nspawn command line.
4729 Contributions from: Cristian Rodríguez, Daniel Mack, David
4730 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
4731 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
4732 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
4733 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
4734 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
4735 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
4736 Andersen, Tom Gundersen, Torstein Husebø
4738 — Berlin, 2015-09-08
4742 * machinectl gained a new verb 'shell' which opens a fresh
4743 shell on the target container or the host. It is similar to
4744 the existing 'login' command of machinectl, but spawns the
4745 shell directly without prompting for username or
4746 password. The pseudo machine '.host' now refers to the local
4747 host and is used by default. Hence, 'machinectl shell' can
4748 be used as replacement for 'su -' which spawns a session as
4749 a fresh systemd unit in a way that is fully isolated from
4750 the originating session.
4752 * systemd-networkd learned to cope with private-zone DHCP
4753 options and allows other programs to query the values.
4755 * SELinux access control when enabling/disabling units is no
4756 longer enforced with this release. The previous implementation
4757 was incorrect, and a new corrected implementation is not yet
4758 available. As unit file operations are still protected via
4759 polkit and D-Bus policy this is not a security problem. Yet,
4760 distributions which care about optimal SELinux support should
4761 probably not stabilize on this release.
4763 * sd-bus gained support for matches of type "arg0has=", that
4764 test for membership of strings in string arrays sent in bus
4767 * systemd-resolved now dumps the contents of its DNS and LLMNR
4768 caches to the logs on reception of the SIGUSR1 signal. This
4769 is useful to debug DNS behaviour.
4771 * The coredumpctl tool gained a new --directory= option to
4772 operate on journal files in a specific directory.
4774 * "systemctl reboot" and related commands gained a new
4775 "--message=" option which may be used to set a free-text
4776 wall message when shutting down or rebooting the
4777 system. This message is also logged, which is useful for
4778 figuring out the reason for a reboot or shutdown a
4781 * The "systemd-resolve-host" tool's -i switch now takes
4782 network interface numbers as alternative to interface names.
4784 * A new unit file setting for services has been introduced:
4785 UtmpMode= allows configuration of how precisely systemd
4786 handles utmp and wtmp entries for the service if this is
4787 enabled. This allows writing services that appear similar to
4788 user sessions in the output of the "w", "who", "last" and
4791 * systemd-resolved will now locally synthesize DNS resource
4792 records for the "localhost" and "gateway" domains as well as
4793 the local hostname. This should ensure that clients querying
4794 RRs via resolved will get similar results as those going via
4795 NSS, if nss-myhostname is enabled.
4797 Contributions from: Alastair Hughes, Alex Crawford, Daniel
4798 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
4799 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
4800 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
4801 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
4802 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
4803 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
4804 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
4805 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
4806 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
4807 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
4808 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4810 — Berlin, 2015-08-27
4814 * The systemd-efi-boot-generator functionality was merged into
4815 systemd-gpt-auto-generator.
4817 * systemd-networkd now supports Group Policy for vxlan
4818 devices. It can be enabled via the new boolean configuration
4819 option called 'GroupPolicyExtension='.
4821 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
4822 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
4823 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
4825 — Berlin, 2015-07-31
4829 * The python-systemd code has been removed from the systemd repository.
4830 A new repository has been created which accommodates the code from
4831 now on, and we kindly ask distributions to create a separate package
4832 for this: https://github.com/systemd/python-systemd
4834 * The systemd daemon will now reload its main configuration
4835 (/etc/systemd/system.conf) on daemon-reload.
4837 * sd-dhcp now exposes vendor specific extensions via
4838 sd_dhcp_lease_get_vendor_specific().
4840 * systemd-networkd gained a number of new configuration options.
4842 - A new boolean configuration option for TAP devices called
4843 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
4844 device, thus allowing to send and receive GSO packets.
4846 - A new tunnel configuration option called 'CopyDSCP='.
4847 If enabled, the DSCP field of ip6 tunnels is copied into the
4848 decapsulated packet.
4850 - A set of boolean bridge configuration options were added.
4851 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
4852 and 'UnicastFlood=' are now parsed by networkd and applied to the
4853 respective bridge link device via the respective IFLA_BRPORT_*
4856 - A new string configuration option to override the hostname sent
4857 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
4858 is true, networkd will use the configured hostname instead of the
4859 system hostname when sending DHCP requests.
4861 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
4862 networkd will configure the IPv6 flow-label of the tunnel device
4863 according to RFC2460.
4865 - The 'macvtap' virtual network devices are now supported, similar to
4866 the already supported 'macvlan' devices.
4868 * systemd-resolved now implements RFC5452 to improve resilience against
4869 cache poisoning. Additionally, source port randomization is enabled
4870 by default to further protect against DNS spoofing attacks.
4872 * nss-mymachines now supports translating UIDs and GIDs of running
4873 containers with user-namespaces enabled. If a container 'foo'
4874 translates a host uid 'UID' to the container uid 'TUID', then
4875 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
4876 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
4877 mapped as 'vg-foo-TGID'.
4879 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
4880 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
4881 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
4882 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
4883 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
4884 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
4885 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
4886 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
4887 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
4888 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
4890 — Berlin, 2015-07-29
4894 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
4895 There are no known issues with current sysfs, and udev does not need
4896 or should be used to work around such bugs.
4898 * udev does no longer enable USB HID power management. Several reports
4899 indicate, that some devices cannot handle that setting.
4901 * The udev accelerometer helper was removed. The functionality
4902 is now fully included in iio-sensor-proxy. But this means,
4903 older iio-sensor-proxy versions will no longer provide
4904 accelerometer/orientation data with this systemd version.
4905 Please upgrade iio-sensor-proxy to version 1.0.
4907 * networkd gained a new configuration option IPv6PrivacyExtensions=
4908 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
4909 for Stateless Address") on selected networks.
4911 * For the sake of fewer build-time dependencies and less code in the
4912 main repository, the python bindings are about to be removed in the
4913 next release. A new repository has been created which accommodates
4914 the code from now on, and we kindly ask distributions to create a
4915 separate package for this. The removal will take place in v223.
4917 https://github.com/systemd/python-systemd
4919 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
4920 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
4921 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
4922 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
4923 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
4924 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
4925 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
4926 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
4927 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4928 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4930 — Berlin, 2015-07-07
4934 * The sd-bus.h and sd-event.h APIs have now been declared
4935 stable and have been added to the official interface of
4936 libsystemd.so. sd-bus implements an alternative D-Bus client
4937 library, that is relatively easy to use, very efficient and
4938 supports both classic D-Bus as well as kdbus as transport
4939 backend. sd-event is a generic event loop abstraction that
4940 is built around Linux epoll, but adds features such as event
4941 prioritization or efficient timer handling. Both APIs are good
4942 choices for C programs looking for a bus and/or event loop
4943 implementation that is minimal and does not have to be
4944 portable to other kernels.
4946 * kdbus support is no longer compile-time optional. It is now
4947 always built-in. However, it can still be disabled at
4948 runtime using the kdbus=0 kernel command line setting, and
4949 that setting may be changed to default to off, by specifying
4950 --disable-kdbus at build-time. Note though that the kernel
4951 command line setting has no effect if the kdbus.ko kernel
4952 module is not installed, in which case kdbus is (obviously)
4953 also disabled. We encourage all downstream distributions to
4954 begin testing kdbus by adding it to the kernel images in the
4955 development distributions, and leaving kdbus support in
4958 * The minimal required util-linux version has been bumped to
4961 * Support for chkconfig (--enable-chkconfig) was removed in
4962 favor of calling an abstraction tool
4963 /lib/systemd/systemd-sysv-install. This needs to be
4964 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
4965 in README for details.
4967 * If there's a systemd unit and a SysV init script for the
4968 same service name, and the user executes "systemctl enable"
4969 for it (or a related call), then this will now enable both
4970 (or execute the related operation on both), not just the
4973 * The libudev API documentation has been converted from gtkdoc
4976 * gudev has been removed from the systemd tree, it is now an
4979 * The systemd-cgtop tool learnt a new --raw switch to generate
4980 "raw" (machine parsable) output.
4982 * networkd's IPForwarding= .network file setting learnt the
4983 new setting "kernel", which ensures that networkd does not
4984 change the IP forwarding sysctl from the default kernel
4987 * The systemd-logind bus API now exposes a new boolean
4988 property "Docked" that reports whether logind considers the
4989 system "docked", i.e. connected to a docking station or not.
4991 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
4992 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
4993 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
4994 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
4995 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
4996 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
4997 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
4998 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
4999 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
5000 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
5001 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
5002 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
5003 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
5004 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
5005 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
5006 Fink, Zbigniew Jędrzejewski-Szmek
5008 — Berlin, 2015-06-19
5012 * The gudev library has been extracted into a separate repository
5013 available at: https://git.gnome.org/browse/libgudev/
5014 It is now managed as part of the Gnome project. Distributions
5015 are recommended to pass --disable-gudev to systemd and use
5016 gudev from the Gnome project instead. gudev is still included
5017 in systemd, for now. It will be removed soon, though. Please
5018 also see the announcement-thread on systemd-devel:
5019 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
5021 * systemd now exposes a CPUUsageNSec= property for each
5022 service unit on the bus, that contains the overall consumed
5023 CPU time of a service (the sum of what each process of the
5024 service consumed). This value is only available if
5025 CPUAccounting= is turned on for a service, and is then shown
5026 in the "systemctl status" output.
5028 * Support for configuring alternative mappings of the old SysV
5029 runlevels to systemd targets has been removed. They are now
5030 hardcoded in a way that runlevels 2, 3, 4 all map to
5031 multi-user.target and 5 to graphical.target (which
5032 previously was already the default behaviour).
5034 * The auto-mounter logic gained support for mount point
5035 expiry, using a new TimeoutIdleSec= setting in .automount
5036 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
5038 * The EFI System Partition (ESP) as mounted to /boot by
5039 systemd-efi-boot-generator will now be unmounted
5040 automatically after 2 minutes of not being used. This should
5041 minimize the risk of ESP corruptions.
5043 * New /etc/fstab options x-systemd.requires= and
5044 x-systemd.requires-mounts-for= are now supported to express
5045 additional dependencies for mounts. This is useful for
5046 journalling file systems that support external journal
5047 devices or overlay file systems that require underlying file
5048 systems to be mounted.
5050 * systemd does not support direct live-upgrades (via systemctl
5051 daemon-reexec) from versions older than v44 anymore. As no
5052 distribution we are aware of shipped such old versions in a
5053 stable release this should not be problematic.
5055 * When systemd forks off a new per-connection service instance
5056 it will now set the $REMOTE_ADDR environment variable to the
5057 remote IP address, and $REMOTE_PORT environment variable to
5058 the remote IP port. This behaviour is similar to the
5059 corresponding environment variables defined by CGI.
5061 * systemd-networkd gained support for uplink failure
5062 detection. The BindCarrier= option allows binding interface
5063 configuration dynamically to the link sense of other
5064 interfaces. This is useful to achieve behaviour like in
5067 * systemd-networkd gained support for configuring the DHCP
5068 client identifier to use when requesting leases.
5070 * systemd-networkd now has a per-network UseNTP= option to
5071 configure whether NTP server information acquired via DHCP
5072 is passed on to services like systemd-timesyncd.
5074 * systemd-networkd gained support for vti6 tunnels.
5076 * Note that systemd-networkd manages the sysctl variable
5077 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
5078 it is configured for since v219. The variable controls IP
5079 forwarding, and is a per-interface alternative to the global
5080 /proc/sys/net/ipv[46]/ip_forward. This setting is
5081 configurable in the IPForward= option, which defaults to
5082 "no". This means if networkd is used for an interface it is
5083 no longer sufficient to set the global sysctl option to turn
5084 on IP forwarding! Instead, the .network file option
5085 IPForward= needs to be turned on! Note that the
5086 implementation of this behaviour was broken in v219 and has
5089 * Many bonding and vxlan options are now configurable in
5092 * systemd-nspawn gained a new --property= setting to set unit
5093 properties for the container scope. This is useful for
5094 setting resource parameters (e.g. "CPUShares=500") on
5095 containers started from the command line.
5097 * systemd-nspawn gained a new --private-users= switch to make
5098 use of user namespacing available on recent Linux kernels.
5100 * systemd-nspawn may now be called as part of a shell pipeline
5101 in which case the pipes used for stdin and stdout are passed
5102 directly to the process invoked in the container, without
5103 indirection via a pseudo tty.
5105 * systemd-nspawn gained a new switch to control the UNIX
5106 signal to use when killing the init process of the container
5109 * systemd-nspawn gained a new --overlay= switch for mounting
5110 overlay file systems into the container using the new kernel
5113 * When a container image is imported via systemd-importd and
5114 the host file system is not btrfs, a loopback block device
5115 file is created in /var/lib/machines.raw with a btrfs file
5116 system inside. It is then mounted to /var/lib/machines to
5117 enable btrfs features for container management. The loopback
5118 file and btrfs file system is grown as needed when container
5119 images are imported via systemd-importd.
5121 * systemd-machined/systemd-importd gained support for btrfs
5122 quota, to enforce container disk space limits on disk. This
5123 is exposed in "machinectl set-limit".
5125 * systemd-importd now can import containers from local .tar,
5126 .raw and .qcow2 images, and export them to .tar and .raw. It
5127 can also import dkr v2 images now from the network (on top
5130 * systemd-importd gained support for verifying downloaded
5131 images with gpg2 (previously only gpg1 was supported).
5133 * systemd-machined, systemd-logind, systemd: most bus calls are
5134 now accessible to unprivileged processes via polkit. Also,
5135 systemd-logind will now allow users to kill their own sessions
5136 without further privileges or authorization.
5138 * systemd-shutdownd has been removed. This service was
5139 previously responsible for implementing scheduled shutdowns
5140 as exposed in /usr/bin/shutdown's time parameter. This
5141 functionality has now been moved into systemd-logind and is
5142 accessible via a bus interface.
5144 * "systemctl reboot" gained a new switch --firmware-setup that
5145 can be used to reboot into the EFI firmware setup, if that
5146 is available. systemd-logind now exposes an API on the bus
5147 to trigger such reboots, in case graphical desktop UIs want
5148 to cover this functionality.
5150 * "systemctl enable", "systemctl disable" and "systemctl mask"
5151 now support a new "--now" switch. If specified the units
5152 that are enabled will also be started, and the ones
5153 disabled/masked also stopped.
5155 * The Gummiboot EFI boot loader tool has been merged into
5156 systemd, and renamed to "systemd-boot". The bootctl tool has been
5157 updated to support systemd-boot.
5159 * An EFI kernel stub has been added that may be used to create
5160 kernel EFI binaries that contain not only the actual kernel,
5161 but also an initrd, boot splash, command line and OS release
5162 information. This combined binary can then be signed as a
5163 single image, so that the firmware can verify it all in one
5164 step. systemd-boot has special support for EFI binaries created
5165 like this and can extract OS release information from them
5166 and show them in the boot menu. This functionality is useful
5167 to implement cryptographically verified boot schemes.
5169 * Optional support has been added to systemd-fsck to pass
5170 fsck's progress report to an AF_UNIX socket in the file
5173 * udev will no longer create device symlinks for all block
5174 devices by default. A blacklist for excluding special block
5175 devices from this logic has been turned into a whitelist
5176 that requires picking block devices explicitly that require
5179 * A new (currently still internal) API sd-device.h has been
5180 added to libsystemd. This modernized API is supposed to
5181 replace libudev eventually. In fact, already much of libudev
5182 is now just a wrapper around sd-device.h.
5184 * A new hwdb database for storing metadata about pointing
5185 stick devices has been added.
5187 * systemd-tmpfiles gained support for setting file attributes
5188 similar to the "chattr" tool with new 'h' and 'H' lines.
5190 * systemd-journald will no longer unconditionally set the
5191 btrfs NOCOW flag on new journal files. This is instead done
5192 with tmpfiles snippet using the new 'h' line type. This
5193 allows easy disabling of this logic, by masking the
5194 journal-nocow.conf tmpfiles file.
5196 * systemd-journald will now translate audit message types to
5197 human readable identifiers when writing them to the
5198 journal. This should improve readability of audit messages.
5200 * The LUKS logic gained support for the offset= and skip=
5201 options in /etc/crypttab, as previously implemented by
5204 * /usr/lib/os-release gained a new optional field VARIANT= for
5205 distributions that support multiple variants (such as a
5206 desktop edition, a server edition, ...)
5208 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
5209 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
5210 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
5211 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
5212 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
5213 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
5214 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
5215 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
5216 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
5217 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
5218 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
5219 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
5220 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
5221 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
5222 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
5223 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
5224 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
5225 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
5226 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
5227 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
5228 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
5229 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
5230 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
5231 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
5232 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
5233 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
5234 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
5236 — Berlin, 2015-05-22
5240 * Introduce a new API "sd-hwdb.h" for querying the hardware
5241 metadata database. With this minimal interface one can query
5242 and enumerate the udev hwdb, decoupled from the old libudev
5243 library. libudev's interface for this is now only a wrapper
5244 around sd-hwdb. A new tool systemd-hwdb has been added to
5245 interface with and update the database.
5247 * When any of systemd's tools copies files (for example due to
5248 tmpfiles' C lines) a btrfs reflink will attempted first,
5249 before bytewise copying is done.
5251 * systemd-nspawn gained a new --ephemeral switch. When
5252 specified a btrfs snapshot is taken of the container's root
5253 directory, and immediately removed when the container
5254 terminates again. Thus, a container can be started whose
5255 changes never alter the container's root directory, and are
5256 lost on container termination. This switch can also be used
5257 for starting a container off the root file system of the
5258 host without affecting the host OS. This switch is only
5259 available on btrfs file systems.
5261 * systemd-nspawn gained a new --template= switch. It takes the
5262 path to a container tree to use as template for the tree
5263 specified via --directory=, should that directory be
5264 missing. This allows instantiating containers dynamically,
5265 on first run. This switch is only available on btrfs file
5268 * When a .mount unit refers to a mount point on which multiple
5269 mounts are stacked, and the .mount unit is stopped all of
5270 the stacked mount points will now be unmounted until no
5271 mount point remains.
5273 * systemd now has an explicit notion of supported and
5274 unsupported unit types. Jobs enqueued for unsupported unit
5275 types will now fail with an "unsupported" error code. More
5276 specifically .swap, .automount and .device units are not
5277 supported in containers, .busname units are not supported on
5278 non-kdbus systems. .swap and .automount are also not
5279 supported if their respective kernel compile time options
5282 * machinectl gained support for two new "copy-from" and
5283 "copy-to" commands for copying files from a running
5284 container to the host or vice versa.
5286 * machinectl gained support for a new "bind" command to bind
5287 mount host directories into local containers. This is
5288 currently only supported for nspawn containers.
5290 * networkd gained support for configuring bridge forwarding
5291 database entries (fdb) from .network files.
5293 * A new tiny daemon "systemd-importd" has been added that can
5294 download container images in tar, raw, qcow2 or dkr formats,
5295 and make them available locally in /var/lib/machines, so
5296 that they can run as nspawn containers. The daemon can GPG
5297 verify the downloads (not supported for dkr, since it has no
5298 provisions for verifying downloads). It will transparently
5299 decompress bz2, xz, gzip compressed downloads if necessary,
5300 and restore sparse files on disk. The daemon uses privilege
5301 separation to ensure the actual download logic runs with
5302 fewer privileges than the daemon itself. machinectl has
5303 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
5304 make the functionality of importd available to the
5305 user. With this in place the Fedora and Ubuntu "Cloud"
5306 images can be downloaded and booted as containers unmodified
5307 (the Fedora images lack the appropriate GPG signature files
5308 currently, so they cannot be verified, but this will change
5309 soon, hopefully). Note that downloading images is currently
5310 only fully supported on btrfs.
5312 * machinectl is now able to list container images found in
5313 /var/lib/machines, along with some metadata about sizes of
5314 disk and similar. If the directory is located on btrfs and
5315 quota is enabled, this includes quota display. A new command
5316 "image-status" has been added that shows additional
5317 information about images.
5319 * machinectl is now able to clone container images
5320 efficiently, if the underlying file system (btrfs) supports
5321 it, with the new "machinectl clone" command. It also
5322 gained commands for renaming and removing images, as well as
5323 marking them read-only or read-write (supported also on
5324 legacy file systems).
5326 * networkd gained support for collecting LLDP network
5327 announcements, from hardware that supports this. This is
5328 shown in networkctl output.
5330 * systemd-run gained support for a new -t (--pty) switch for
5331 invoking a binary on a pty whose input and output is
5332 connected to the invoking terminal. This allows executing
5333 processes as system services while interactively
5334 communicating with them via the terminal. Most interestingly
5335 this is supported across container boundaries. Invoking
5336 "systemd-run -t /bin/bash" is an alternative to running a
5337 full login session, the difference being that the former
5338 will not register a session, nor go through the PAM session
5341 * tmpfiles gained support for a new "v" line type for creating
5342 btrfs subvolumes. If the underlying file system is a legacy
5343 file system, this automatically degrades to creating a
5344 normal directory. Among others /var/lib/machines is now
5345 created like this at boot, should it be missing.
5347 * The directory /var/lib/containers/ has been deprecated and
5348 been replaced by /var/lib/machines. The term "machines" has
5349 been used in the systemd context as generic term for both
5350 VMs and containers, and hence appears more appropriate for
5351 this, as the directory can also contain raw images bootable
5354 * systemd-nspawn when invoked with -M but without --directory=
5355 or --image= is now capable of searching for the container
5356 root directory, subvolume or disk image automatically, in
5357 /var/lib/machines. systemd-nspawn@.service has been updated
5358 to make use of this, thus allowing it to be used for raw
5361 * A new machines.target unit has been introduced that is
5362 supposed to group all containers/VMs invoked as services on
5363 the system. systemd-nspawn@.service has been updated to
5364 integrate with that.
5366 * machinectl gained a new "start" command, for invoking a
5367 container as a service. "machinectl start foo" is mostly
5368 equivalent to "systemctl start systemd-nspawn@foo.service",
5369 but handles escaping in a nicer way.
5371 * systemd-nspawn will now mount most of the cgroupfs tree
5372 read-only into each container, with the exception of the
5373 container's own subtree in the name=systemd hierarchy.
5375 * journald now sets the special FS_NOCOW file flag for its
5376 journal files. This should improve performance on btrfs, by
5377 avoiding heavy fragmentation when journald's write-pattern
5378 is used on COW file systems. It degrades btrfs' data
5379 integrity guarantees for the files to the same levels as for
5380 ext3/ext4 however. This should be OK though as journald does
5381 its own data integrity checks and all its objects are
5382 checksummed on disk. Also, journald should handle btrfs disk
5383 full events a lot more gracefully now, by processing SIGBUS
5384 errors, and not relying on fallocate() anymore.
5386 * When journald detects that journal files it is writing to
5387 have been deleted it will immediately start new journal
5390 * systemd now provides a way to store file descriptors
5391 per-service in PID 1. This is useful for daemons to ensure
5392 that fds they require are not lost during a daemon
5393 restart. The fds are passed to the daemon on the next
5394 invocation in the same way socket activation fds are
5395 passed. This is now used by journald to ensure that the
5396 various sockets connected to all the system's stdout/stderr
5397 are not lost when journald is restarted. File descriptors
5398 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
5399 an extension to sd_notify(). Note that a limit is enforced
5400 on the number of fds a service can store in PID 1, and it
5401 defaults to 0, so that no fds may be stored, unless this is
5402 explicitly turned on.
5404 * The default TERM variable to use for units connected to a
5405 terminal, when no other value is explicitly is set is now
5406 vt220 rather than vt102. This should be fairly safe still,
5407 but allows PgUp/PgDn work.
5409 * The /etc/crypttab option header= as known from Debian is now
5412 * "loginctl user-status" and "loginctl session-status" will
5413 now show the last 10 lines of log messages of the
5414 user/session following the status output. Similar,
5415 "machinectl status" will show the last 10 log lines
5416 associated with a virtual machine or container
5417 service. (Note that this is usually not the log messages
5418 done in the VM/container itself, but simply what the
5419 container manager logs. For nspawn this includes all console
5422 * "loginctl session-status" without further argument will now
5423 show the status of the session of the caller. Similar,
5424 "lock-session", "unlock-session", "activate",
5425 "enable-linger", "disable-linger" may now be called without
5426 session/user parameter in which case they apply to the
5427 caller's session/user.
5429 * An X11 session scriptlet is now shipped that uploads
5430 $DISPLAY and $XAUTHORITY into the environment of the systemd
5431 --user daemon if a session begins. This should improve
5432 compatibility with X11 enabled applications run as systemd
5435 * Generators are now subject to masking via /etc and /run, the
5436 same way as unit files.
5438 * networkd .network files gained support for configuring
5439 per-link IPv4/IPv6 packet forwarding as well as IPv4
5440 masquerading. This is by default turned on for veth links to
5441 containers, as registered by systemd-nspawn. This means that
5442 nspawn containers run with --network-veth will now get
5443 automatic routed access to the host's networks without any
5444 further configuration or setup, as long as networkd runs on
5447 * systemd-nspawn gained the --port= (-p) switch to expose TCP
5448 or UDP posts of a container on the host. With this in place
5449 it is possible to run containers with private veth links
5450 (--network-veth), and have their functionality exposed on
5451 the host as if their services were running directly on the
5454 * systemd-nspawn's --network-veth switch now gained a short
5455 version "-n", since with the changes above it is now truly
5456 useful out-of-the-box. The systemd-nspawn@.service has been
5457 updated to make use of it too by default.
5459 * systemd-nspawn will now maintain a per-image R/W lock, to
5460 ensure that the same image is not started more than once
5461 writable. (It's OK to run an image multiple times
5462 simultaneously in read-only mode.)
5464 * systemd-nspawn's --image= option is now capable of
5465 dissecting and booting MBR and GPT disk images that contain
5466 only a single active Linux partition. Previously it
5467 supported only GPT disk images with proper GPT type
5468 IDs. This allows running cloud images from major
5469 distributions directly with systemd-nspawn, without
5472 * In addition to collecting mouse dpi data in the udev
5473 hardware database, there's now support for collecting angle
5474 information for mouse scroll wheels. The database is
5475 supposed to guarantee similar scrolling behavior on mice
5476 that it knows about. There's also support for collecting
5477 information about Touchpad types.
5479 * udev's input_id built-in will now also collect touch screen
5480 dimension data and attach it to probed devices.
5482 * /etc/os-release gained support for a Distribution Privacy
5485 * networkd gained support for creating "ipvlan", "gretap",
5486 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
5488 * systemd-tmpfiles gained support for "a" lines for setting
5491 * systemd-nspawn will now mount /tmp in the container to
5492 tmpfs, automatically.
5494 * systemd now exposes the memory.usage_in_bytes cgroup
5495 attribute and shows it for each service in the "systemctl
5496 status" output, if available.
5498 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
5499 immediate reboot is triggered. This useful if shutdown is
5500 hung and is unable to complete, to expedite the
5501 operation. Note that this kind of reboot will still unmount
5502 all file systems, and hence should not result in fsck being
5505 * A .device unit for an optical block device will now be
5506 considered active only when a medium is in the drive. Also,
5507 mount units are now bound to their backing devices thus
5508 triggering automatic unmounting when devices become
5509 unavailable. With this in place systemd will now
5510 automatically unmount left-over mounts when a CD-ROM is
5511 ejected or an USB stick is yanked from the system.
5513 * networkd-wait-online now has support for waiting for
5514 specific interfaces only (with globbing), and for giving up
5515 after a configurable timeout.
5517 * networkd now exits when idle. It will be automatically
5518 restarted as soon as interfaces show up, are removed or
5519 change state. networkd will stay around as long as there is
5520 at least one DHCP state machine or similar around, that keep
5523 * networkd may now configure IPv6 link-local addressing in
5524 addition to IPv4 link-local addressing.
5526 * The IPv6 "token" for use in SLAAC may now be configured for
5527 each .network interface in networkd.
5529 * Routes configured with networkd may now be assigned a scope
5532 * networkd's [Match] sections now support globbing and lists
5533 of multiple space-separated matches per item.
5535 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
5536 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
5537 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
5538 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
5539 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
5540 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
5541 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
5542 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
5543 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
5544 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
5545 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
5546 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
5547 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
5548 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
5549 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
5550 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
5551 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
5552 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
5553 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
5554 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
5555 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
5556 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
5557 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
5558 Hoffmann, Zbigniew Jędrzejewski-Szmek
5560 — Berlin, 2015-02-16
5564 * When querying unit file enablement status (for example via
5565 "systemctl is-enabled"), a new state "indirect" is now known
5566 which indicates that a unit might not be enabled itself, but
5567 another unit listed in its Also= setting might be.
5569 * Similar to the various existing ConditionXYZ= settings for
5570 units, there are now matching AssertXYZ= settings. While
5571 failing conditions cause a unit to be skipped, but its job
5572 to succeed, failing assertions declared like this will cause
5573 a unit start operation and its job to fail.
5575 * hostnamed now knows a new chassis type "embedded".
5577 * systemctl gained a new "edit" command. When used on a unit
5578 file, this allows extending unit files with .d/ drop-in
5579 configuration snippets or editing the full file (after
5580 copying it from /usr/lib to /etc). This will invoke the
5581 user's editor (as configured with $EDITOR), and reload the
5582 modified configuration after editing.
5584 * "systemctl status" now shows the suggested enablement state
5585 for a unit, as declared in the (usually vendor-supplied)
5586 system preset files.
5588 * nss-myhostname will now resolve the single-label host name
5589 "gateway" to the locally configured default IP routing
5590 gateways, ordered by their metrics. This assigns a stable
5591 name to the used gateways, regardless which ones are
5592 currently configured. Note that the name will only be
5593 resolved after all other name sources (if nss-myhostname is
5594 configured properly) and should hence not negatively impact
5595 systems that use the single-label host name "gateway" in
5598 * systemd-inhibit now allows filtering by mode when listing
5601 * Scope and service units gained a new "Delegate" boolean
5602 property, which, when set, allows processes running inside the
5603 unit to further partition resources. This is primarily
5604 useful for systemd user instances as well as container
5607 * journald will now pick up audit messages directly from
5608 the kernel, and log them like any other log message. The
5609 audit fields are split up and fully indexed. This means that
5610 journalctl in many ways is now a (nicer!) alternative to
5611 ausearch, the traditional audit client. Note that this
5612 implements only a minimal audit client. If you want the
5613 special audit modes like reboot-on-log-overflow, please use
5614 the traditional auditd instead, which can be used in
5615 parallel to journald.
5617 * The ConditionSecurity= unit file option now understands the
5618 special string "audit" to check whether auditing is
5621 * journalctl gained two new commands --vacuum-size= and
5622 --vacuum-time= to delete old journal files until the
5623 remaining ones take up no more than the specified size on disk,
5624 or are not older than the specified time.
5626 * A new, native PPPoE library has been added to sd-network,
5627 systemd's library of light-weight networking protocols. This
5628 library will be used in a future version of networkd to
5629 enable PPPoE communication without an external pppd daemon.
5631 * The busctl tool now understands a new "capture" verb that
5632 works similar to "monitor", but writes a packet capture
5633 trace to STDOUT that can be redirected to a file which is
5634 compatible with libcap's capture file format. This can then
5635 be loaded in Wireshark and similar tools to inspect bus
5638 * The busctl tool now understands a new "tree" verb that shows
5639 the object trees of a specific service on the bus, or of all
5642 * The busctl tool now understands a new "introspect" verb that
5643 shows all interfaces and members of objects on the bus,
5644 including their signature and values. This is particularly
5645 useful to get more information about bus objects shown by
5646 the new "busctl tree" command.
5648 * The busctl tool now understands new verbs "call",
5649 "set-property" and "get-property" for invoking bus method
5650 calls, setting and getting bus object properties in a
5653 * busctl gained a new --augment-creds= argument that controls
5654 whether the tool shall augment credential information it
5655 gets from the bus with data from /proc, in a possibly
5658 * nspawn's --link-journal= switch gained two new values
5659 "try-guest" and "try-host" that work like "guest" and
5660 "host", but do not fail if the host has no persistent
5661 journalling enabled. -j is now equivalent to
5662 --link-journal=try-guest.
5664 * macvlan network devices created by nspawn will now have
5665 stable MAC addresses.
5667 * A new SmackProcessLabel= unit setting has been added, which
5668 controls the SMACK security label processes forked off by
5669 the respective unit shall use.
5671 * If compiled with --enable-xkbcommon, systemd-localed will
5672 verify x11 keymap settings by compiling the given keymap. It
5673 will spew out warnings if the compilation fails. This
5674 requires libxkbcommon to be installed.
5676 * When a coredump is collected, a larger number of metadata
5677 fields is now collected and included in the journal records
5678 created for it. More specifically, control group membership,
5679 environment variables, memory maps, working directory,
5680 chroot directory, /proc/$PID/status, and a list of open file
5681 descriptors is now stored in the log entry.
5683 * The udev hwdb now contains DPI information for mice. For
5686 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
5688 * All systemd programs that read standalone configuration
5689 files in /etc now also support a corresponding series of
5690 .conf.d configuration directories in /etc/, /run/,
5691 /usr/local/lib/, /usr/lib/, and (if configured with
5692 --enable-split-usr) /lib/. In particular, the following
5693 configuration files now have corresponding configuration
5694 directories: system.conf user.conf, logind.conf,
5695 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
5696 resolved.conf, timesyncd.conf, journal-remote.conf, and
5697 journal-upload.conf. Note that distributions should use the
5698 configuration directories in /usr/lib/; the directories in
5699 /etc/ are reserved for the system administrator.
5701 * systemd-rfkill will no longer take the rfkill device name
5702 into account when storing rfkill state on disk, as the name
5703 might be dynamically assigned and not stable. Instead, the
5704 ID_PATH udev variable combined with the rfkill type (wlan,
5705 bluetooth, ...) is used.
5707 * A new service systemd-machine-id-commit.service has been
5708 added. When used on systems where /etc is read-only during
5709 boot, and /etc/machine-id is not initialized (but an empty
5710 file), this service will copy the temporary machine ID
5711 created as replacement into /etc after the system is fully
5712 booted up. This is useful for systems that are freshly
5713 installed with a non-initialized machine ID, but should get
5714 a fixed machine ID for subsequent boots.
5716 * networkd's .netdev files now provide a large set of
5717 configuration parameters for VXLAN devices. Similarly, the
5718 bridge port cost parameter is now configurable in .network
5719 files. There's also new support for configuring IP source
5720 routing. networkd .link files gained support for a new
5721 OriginalName= match that is useful to match against the
5722 original interface name the kernel assigned. .network files
5723 may include MTU= and MACAddress= fields for altering the MTU
5724 and MAC address while being connected to a specific network
5727 * The LUKS logic gained supported for configuring
5728 UUID-specific key files. There's also new support for naming
5729 LUKS device from the kernel command line, using the new
5730 luks.name= argument.
5732 * Timer units may now be transiently created via the bus API
5733 (this was previously already available for scope and service
5734 units). In addition it is now possible to create multiple
5735 transient units at the same time with a single bus call. The
5736 "systemd-run" tool has been updated to make use of this for
5737 running commands on a specified time, in at(1)-style.
5739 * tmpfiles gained support for "t" lines, for assigning
5740 extended attributes to files. Among other uses this may be
5741 used to assign SMACK labels to files.
5743 Contributions from: Alin Rauta, Alison Chaiken, Andrej
5744 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
5745 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
5746 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
5747 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
5748 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
5749 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
5750 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
5751 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
5752 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
5753 Michael Biebl, Michael Chapman, Michael Marineau, Michal
5754 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
5755 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
5756 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
5757 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
5758 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5759 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
5760 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
5762 — Berlin, 2014-12-10
5766 * journalctl gained the new options -t/--identifier= to match
5767 on the syslog identifier (aka "tag"), as well as --utc to
5768 show log timestamps in the UTC timezone. journalctl now also
5769 accepts -n/--lines=all to disable line capping in a pager.
5771 * journalctl gained a new switch, --flush, that synchronously
5772 flushes logs from /run/log/journal to /var/log/journal if
5773 persistent storage is enabled. systemd-journal-flush.service
5774 now waits until the operation is complete.
5776 * Services can notify the manager before they start a reload
5777 (by sending RELOADING=1) or shutdown (by sending
5778 STOPPING=1). This allows the manager to track and show the
5779 internal state of daemons and closes a race condition when
5780 the process is still running but has closed its D-Bus
5783 * Services with Type=oneshot do not have to have any ExecStart
5786 * User units are now loaded also from
5787 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
5788 /run/systemd/user directory that was already previously
5789 supported, but is under the control of the user.
5791 * Job timeouts (i.e. timeouts on the time a job that is
5792 queued stays in the run queue) can now optionally result in
5793 immediate reboot or power-off actions (JobTimeoutAction= and
5794 JobTimeoutRebootArgument=). This is useful on ".target"
5795 units, to limit the maximum time a target remains
5796 undispatched in the run queue, and to trigger an emergency
5797 operation in such a case. This is now used by default to
5798 turn off the system if boot-up (as defined by everything in
5799 basic.target) hangs and does not complete for at least
5800 15min. Also, if power-off or reboot hang for at least 30min
5801 an immediate power-off/reboot operation is triggered. This
5802 functionality is particularly useful to increase reliability
5803 on embedded devices, but also on laptops which might
5804 accidentally get powered on when carried in a backpack and
5805 whose boot stays stuck in a hard disk encryption passphrase
5808 * systemd-logind can be configured to also handle lid switch
5809 events even when the machine is docked or multiple displays
5810 are attached (HandleLidSwitchDocked= option).
5812 * A helper binary and a service have been added which can be
5813 used to resume from hibernation in the initramfs. A
5814 generator will parse the resume= option on the kernel
5815 command line to trigger resume.
5817 * A user console daemon systemd-consoled has been
5818 added. Currently, it is a preview, and will so far open a
5819 single terminal on each session of the user marked as
5820 Desktop=systemd-console.
5822 * Route metrics can be specified for DHCP routes added by
5825 * The SELinux context of socket-activated services can be set
5826 from the information provided by the networking stack
5827 (SELinuxContextFromNet= option).
5829 * Userspace firmware loading support has been removed and
5830 the minimum supported kernel version is thus bumped to 3.7.
5832 * Timeout for udev workers has been increased from 1 to 3
5833 minutes, but a warning will be printed after 1 minute to
5834 help diagnose kernel modules that take a long time to load.
5836 * Udev rules can now remove tags on devices with TAG-="foobar".
5838 * systemd's readahead implementation has been removed. In many
5839 circumstances it didn't give expected benefits even for
5840 rotational disk drives and was becoming less relevant in the
5841 age of SSDs. As none of the developers has been using
5842 rotating media anymore, and nobody stepped up to actively
5843 maintain this component of systemd it has now been removed.
5845 * Swap units can use Options= to specify discard options.
5846 Discard options specified for swaps in /etc/fstab are now
5849 * Docker containers are now detected as a separate type of
5852 * The Password Agent protocol gained support for queries where
5853 the user input is shown, useful e.g. for user names.
5854 systemd-ask-password gained a new --echo option to turn that
5857 * The default sysctl.d/ snippets will now set:
5859 net.core.default_qdisc = fq_codel
5861 This selects Fair Queuing Controlled Delay as the default
5862 queuing discipline for network interfaces. fq_codel helps
5863 fight the network bufferbloat problem. It is believed to be
5864 a good default with no tuning required for most workloads.
5865 Downstream distributions may override this choice. On 10Gbit
5866 servers that do not do forwarding, "fq" may perform better.
5867 Systems without a good clocksource should use "pfifo_fast".
5869 * If kdbus is enabled during build a new option BusPolicy= is
5870 available for service units, that allows locking all service
5871 processes into a stricter bus policy, in order to limit
5872 access to various bus services, or even hide most of them
5873 from the service's view entirely.
5875 * networkctl will now show the .network and .link file
5876 networkd has applied to a specific interface.
5878 * sd-login gained a new API call sd_session_get_desktop() to
5879 query which desktop environment has been selected for a
5882 * UNIX utmp support is now compile-time optional to support
5883 legacy-free systems.
5885 * systemctl gained two new commands "add-wants" and
5886 "add-requires" for pulling in units from specific targets
5889 * If the word "rescue" is specified on the kernel command line
5890 the system will now boot into rescue mode (aka
5891 rescue.target), which was previously available only by
5892 specifying "1" or "systemd.unit=rescue.target" on the kernel
5893 command line. This new kernel command line option nicely
5894 mirrors the already existing "emergency" kernel command line
5897 * New kernel command line options mount.usr=, mount.usrflags=,
5898 mount.usrfstype= have been added that match root=, rootflags=,
5899 rootfstype= but allow mounting a specific file system to
5902 * The $NOTIFY_SOCKET is now also passed to control processes of
5903 services, not only the main process.
5905 * This version reenables support for fsck's -l switch. This
5906 means at least version v2.25 of util-linux is required for
5907 operation, otherwise dead-locks on device nodes may
5908 occur. Again: you need to update util-linux to at least
5909 v2.25 when updating systemd to v217.
5911 * The "multi-seat-x" tool has been removed from systemd, as
5912 its functionality has been integrated into X servers 1.16,
5913 and the tool is hence redundant. It is recommended to update
5914 display managers invoking this tool to simply invoke X
5915 directly from now on, again.
5917 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
5918 message flag has been added for all of systemd's polkit
5919 authenticated method calls has been added. In particular this
5920 now allows optional interactive authorization via polkit for
5921 many of PID1's privileged operations such as unit file
5922 enabling and disabling.
5924 * "udevadm hwdb --update" learnt a new switch "--usr" for
5925 placing the rebuilt hardware database in /usr instead of
5926 /etc. When used only hardware database entries stored in
5927 /usr will be used, and any user database entries in /etc are
5928 ignored. This functionality is useful for vendors to ship a
5929 pre-built database on systems where local configuration is
5930 unnecessary or unlikely.
5932 * Calendar time specifications in .timer units now also
5933 understand the strings "semi-annually", "quarterly" and
5934 "minutely" as shortcuts (in addition to the preexisting
5935 "anually", "hourly", ...).
5937 * systemd-tmpfiles will now correctly create files in /dev
5938 at boot which are marked for creation only at boot. It is
5939 recommended to always create static device nodes with 'c!'
5940 and 'b!', so that they are created only at boot and not
5941 overwritten at runtime.
5943 * When the watchdog logic is used for a service (WatchdogSec=)
5944 and the watchdog timeout is hit the service will now be
5945 terminated with SIGABRT (instead of just SIGTERM), in order
5946 to make sure a proper coredump and backtrace is
5947 generated. This ensures that hanging services will result in
5948 similar coredump/backtrace behaviour as services that hit a
5951 Contributions from: Andreas Henriksson, Andrei Borzenkov,
5952 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
5953 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
5954 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
5955 Herrmann, David Sommerseth, David Strauss, Emil Renner
5956 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
5957 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
5958 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
5959 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
5960 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
5961 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
5962 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
5963 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
5964 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
5965 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
5966 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
5967 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
5968 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
5969 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
5970 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
5971 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
5974 — Berlin, 2014-10-28
5978 * timedated no longer reads NTP implementation unit names from
5979 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
5980 implementations should add a
5982 Conflicts=systemd-timesyncd.service
5984 to their unit files to take over and replace systemd's NTP
5985 default functionality.
5987 * systemd-sysusers gained a new line type "r" for configuring
5988 which UID/GID ranges to allocate system users/groups
5989 from. Lines of type "u" may now add an additional column
5990 that specifies the home directory for the system user to be
5991 created. Also, systemd-sysusers may now optionally read user
5992 information from STDIN instead of a file. This is useful for
5993 invoking it from RPM preinst scriptlets that need to create
5994 users before the first RPM file is installed since these
5995 files might need to be owned by them. A new
5996 %sysusers_create_inline RPM macro has been introduced to do
5997 just that. systemd-sysusers now updates the shadow files as
5998 well as the user/group databases, which should enhance
5999 compatibility with certain tools like grpck.
6001 * A number of bus APIs of PID 1 now optionally consult polkit to
6002 permit access for otherwise unprivileged clients under certain
6003 conditions. Note that this currently doesn't support
6004 interactive authentication yet, but this is expected to be
6005 added eventually, too.
6007 * /etc/machine-info now has new fields for configuring the
6008 deployment environment of the machine, as well as the
6009 location of the machine. hostnamectl has been updated with
6010 new command to update these fields.
6012 * systemd-timesyncd has been updated to automatically acquire
6013 NTP server information from systemd-networkd, which might
6014 have been discovered via DHCP.
6016 * systemd-resolved now includes a caching DNS stub resolver
6017 and a complete LLMNR name resolution implementation. A new
6018 NSS module "nss-resolve" has been added which can be used
6019 instead of glibc's own "nss-dns" to resolve hostnames via
6020 systemd-resolved. Hostnames, addresses and arbitrary RRs may
6021 be resolved via systemd-resolved D-Bus APIs. In contrast to
6022 the glibc internal resolver systemd-resolved is aware of
6023 multi-homed system, and keeps DNS server and caches separate
6024 and per-interface. Queries are sent simultaneously on all
6025 interfaces that have DNS servers configured, in order to
6026 properly handle VPNs and local LANs which might resolve
6027 separate sets of domain names. systemd-resolved may acquire
6028 DNS server information from systemd-networkd automatically,
6029 which in turn might have discovered them via DHCP. A tool
6030 "systemd-resolve-host" has been added that may be used to
6031 query the DNS logic in resolved. systemd-resolved implements
6032 IDNA and automatically uses IDNA or UTF-8 encoding depending
6033 on whether classic DNS or LLMNR is used as transport. In the
6034 next releases we intend to add a DNSSEC and mDNS/DNS-SD
6035 implementation to systemd-resolved.
6037 * A new NSS module nss-mymachines has been added, that
6038 automatically resolves the names of all local registered
6039 containers to their respective IP addresses.
6041 * A new client tool "networkctl" for systemd-networkd has been
6042 added. It currently is entirely passive and will query
6043 networking configuration from udev, rtnetlink and networkd,
6044 and present it to the user in a very friendly
6045 way. Eventually, we hope to extend it to become a full
6046 control utility for networkd.
6048 * .socket units gained a new DeferAcceptSec= setting that
6049 controls the kernels' TCP_DEFER_ACCEPT sockopt for
6050 TCP. Similarly, support for controlling TCP keep-alive
6051 settings has been added (KeepAliveTimeSec=,
6052 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
6053 turning off Nagle's algorithm on TCP has been added
6056 * logind learned a new session type "web", for use in projects
6057 like Cockpit which register web clients as PAM sessions.
6059 * timer units with at least one OnCalendar= setting will now
6060 be started only after time-sync.target has been
6061 reached. This way they will not elapse before the system
6062 clock has been corrected by a local NTP client or
6063 similar. This is particular useful on RTC-less embedded
6064 machines, that come up with an invalid system clock.
6066 * systemd-nspawn's --network-veth= switch should now result in
6067 stable MAC addresses for both the outer and the inner side
6070 * systemd-nspawn gained a new --volatile= switch for running
6071 container instances with /etc or /var unpopulated.
6073 * The kdbus client code has been updated to use the new Linux
6074 3.17 memfd subsystem instead of the old kdbus-specific one.
6076 * systemd-networkd's DHCP client and server now support
6077 FORCERENEW. There are also new configuration options to
6078 configure the vendor client identifier and broadcast mode
6081 * systemd will no longer inform the kernel about the current
6082 timezone, as this is necessarily incorrect and racy as the
6083 kernel has no understanding of DST and similar
6084 concepts. This hence means FAT timestamps will be always
6085 considered UTC, similar to what Android is already
6086 doing. Also, when the RTC is configured to the local time
6087 (rather than UTC) systemd will never synchronize back to it,
6088 as this might confuse Windows at a later boot.
6090 * systemd-analyze gained a new command "verify" for offline
6091 validation of unit files.
6093 * systemd-networkd gained support for a couple of additional
6094 settings for bonding networking setups. Also, the metric for
6095 statically configured routes may now be configured. For
6096 network interfaces where this is appropriate the peer IP
6097 address may now be configured.
6099 * systemd-networkd's DHCP client will no longer request
6100 broadcasting by default, as this tripped up some networks.
6101 For hardware where broadcast is required the feature should
6102 be switched back on using RequestBroadcast=yes.
6104 * systemd-networkd will now set up IPv4LL addresses (when
6105 enabled) even if DHCP is configured successfully.
6107 * udev will now default to respect network device names given
6108 by the kernel when the kernel indicates that these are
6109 predictable. This behavior can be tweaked by changing
6110 NamePolicy= in the relevant .link file.
6112 * A new library systemd-terminal has been added that
6113 implements full TTY stream parsing and rendering. This
6114 library is supposed to be used later on for implementing a
6115 full userspace VT subsystem, replacing the current kernel
6118 * A new tool systemd-journal-upload has been added to push
6119 journal data to a remote system running
6120 systemd-journal-remote.
6122 * journald will no longer forward all local data to another
6123 running syslog daemon. This change has been made because
6124 rsyslog (which appears to be the most commonly used syslog
6125 implementation these days) no longer makes use of this, and
6126 instead pulls the data out of the journal on its own. Since
6127 forwarding the messages to a non-existent syslog server is
6128 more expensive than we assumed we have now turned this
6129 off. If you run a syslog server that is not a recent rsyslog
6130 version, you have to turn this option on again
6131 (ForwardToSyslog= in journald.conf).
6133 * journald now optionally supports the LZ4 compressor for
6134 larger journal fields. This compressor should perform much
6135 better than XZ which was the previous default.
6137 * machinectl now shows the IP addresses of local containers,
6138 if it knows them, plus the interface name of the container.
6140 * A new tool "systemd-escape" has been added that makes it
6141 easy to escape strings to build unit names and similar.
6143 * sd_notify() messages may now include a new ERRNO= field
6144 which is parsed and collected by systemd and shown among the
6145 "systemctl status" output for a service.
6147 * A new component "systemd-firstboot" has been added that
6148 queries the most basic systemd information (timezone,
6149 hostname, root password) interactively on first
6150 boot. Alternatively it may also be used to provision these
6151 things offline on OS images installed into directories.
6153 * The default sysctl.d/ snippets will now set
6155 net.ipv4.conf.default.promote_secondaries=1
6157 This has the benefit of no flushing secondary IP addresses
6158 when primary addresses are removed.
6160 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
6161 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
6162 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
6163 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
6164 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
6165 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
6166 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
6167 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
6168 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
6169 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
6170 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
6171 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
6172 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
6173 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
6174 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
6176 — Berlin, 2014-08-19
6180 * A new tool systemd-sysusers has been added. This tool
6181 creates system users and groups in /etc/passwd and
6182 /etc/group, based on static declarative system user/group
6183 definitions in /usr/lib/sysusers.d/. This is useful to
6184 enable factory resets and volatile systems that boot up with
6185 an empty /etc directory, and thus need system users and
6186 groups created during early boot. systemd now also ships
6187 with two default sysusers.d/ files for the most basic
6188 users and groups systemd and the core operating system
6191 * A new tmpfiles snippet has been added that rebuilds the
6192 essential files in /etc on boot, should they be missing.
6194 * A directive for ensuring automatic clean-up of
6195 /var/cache/man/ has been removed from the default
6196 configuration. This line should now be shipped by the man
6197 implementation. The necessary change has been made to the
6198 man-db implementation. Note that you need to update your man
6199 implementation to one that ships this line, otherwise no
6200 automatic clean-up of /var/cache/man will take place.
6202 * A new condition ConditionNeedsUpdate= has been added that
6203 may conditionalize services to only run when /etc or /var
6204 are "older" than the vendor operating system resources in
6205 /usr. This is useful for reconstructing or updating /etc
6206 after an offline update of /usr or a factory reset, on the
6207 next reboot. Services that want to run once after such an
6208 update or reset should use this condition and order
6209 themselves before the new systemd-update-done.service, which
6210 will mark the two directories as fully updated. A number of
6211 service files have been added making use of this, to rebuild
6212 the udev hardware database, the journald message catalog and
6213 dynamic loader cache (ldconfig). The systemd-sysusers tool
6214 described above also makes use of this now. With this in
6215 place it is now possible to start up a minimal operating
6216 system with /etc empty cleanly. For more information on the
6217 concepts involved see this recent blog story:
6219 http://0pointer.de/blog/projects/stateless.html
6221 * A new system group "input" has been introduced, and all
6222 input device nodes get this group assigned. This is useful
6223 for system-level software to get access to input devices. It
6224 complements what is already done for "audio" and "video".
6226 * systemd-networkd learnt minimal DHCPv4 server support in
6227 addition to the existing DHCPv4 client support. It also
6228 learnt DHCPv6 client and IPv6 Router Solicitation client
6229 support. The DHCPv4 client gained support for static routes
6230 passed in from the server. Note that the [DHCPv4] section
6231 known in older systemd-networkd versions has been renamed to
6232 [DHCP] and is now also used by the DHCPv6 client. Existing
6233 .network files using settings of this section should be
6234 updated, though compatibility is maintained. Optionally, the
6235 client hostname may now be sent to the DHCP server.
6237 * networkd gained support for vxlan virtual networks as well
6238 as tun/tap and dummy devices.
6240 * networkd gained support for automatic allocation of address
6241 ranges for interfaces from a system-wide pool of
6242 addresses. This is useful for dynamically managing a large
6243 number of interfaces with a single network configuration
6244 file. In particular this is useful to easily assign
6245 appropriate IP addresses to the veth links of a large number
6246 of nspawn instances.
6248 * RPM macros for processing sysusers, sysctl and binfmt
6249 drop-in snippets at package installation time have been
6252 * The /etc/os-release file should now be placed in
6253 /usr/lib/os-release. The old location is automatically
6254 created as symlink. /usr/lib is the more appropriate
6255 location of this file, since it shall actually describe the
6256 vendor operating system shipped in /usr, and not the
6257 configuration stored in /etc.
6259 * .mount units gained a new boolean SloppyOptions= setting
6260 that maps to mount(8)'s -s option which enables permissive
6261 parsing of unknown mount options.
6263 * tmpfiles learnt a new "L+" directive which creates a symlink
6264 but (unlike "L") deletes a pre-existing file first, should
6265 it already exist and not already be the correct
6266 symlink. Similarly, "b+", "c+" and "p+" directives have been
6267 added as well, which create block and character devices, as
6268 well as fifos in the filesystem, possibly removing any
6269 pre-existing files of different types.
6271 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
6272 'argument' field (which so far specified the source to
6273 symlink/copy the files from) is now optional. If omitted the
6274 same file os copied from /usr/share/factory/ suffixed by the
6275 full destination path. This is useful for populating /etc
6276 with essential files, by copying them from vendor defaults
6277 shipped in /usr/share/factory/etc.
6279 * A new command "systemctl preset-all" has been added that
6280 applies the service preset settings to all installed unit
6281 files. A new switch --preset-mode= has been added that
6282 controls whether only enable or only disable operations
6285 * A new command "systemctl is-system-running" has been added
6286 that allows checking the overall state of the system, for
6287 example whether it is fully up and running.
6289 * When the system boots up with an empty /etc, the equivalent
6290 to "systemctl preset-all" is executed during early boot, to
6291 make sure all default services are enabled after a factory
6294 * systemd now contains a minimal preset file that enables the
6295 most basic services systemd ships by default.
6297 * Unit files' [Install] section gained a new DefaultInstance=
6298 field for defining the default instance to create if a
6299 template unit is enabled with no instance specified.
6301 * A new passive target cryptsetup-pre.target has been added
6302 that may be used by services that need to make they run and
6303 finish before the first LUKS cryptographic device is set up.
6305 * The /dev/loop-control and /dev/btrfs-control device nodes
6306 are now owned by the "disk" group by default, opening up
6307 access to this group.
6309 * systemd-coredump will now automatically generate a
6310 stack trace of all core dumps taking place on the system,
6311 based on elfutils' libdw library. This stack trace is logged
6314 * systemd-coredump may now optionally store coredumps directly
6315 on disk (in /var/lib/systemd/coredump, possibly compressed),
6316 instead of storing them unconditionally in the journal. This
6317 mode is the new default. A new configuration file
6318 /etc/systemd/coredump.conf has been added to configure this
6319 and other parameters of systemd-coredump.
6321 * coredumpctl gained a new "info" verb to show details about a
6322 specific coredump. A new switch "-1" has also been added
6323 that makes sure to only show information about the most
6324 recent entry instead of all entries. Also, as the tool is
6325 generally useful now the "systemd-" prefix of the binary
6326 name has been removed. Distributions that want to maintain
6327 compatibility with the old name should add a symlink from
6328 the old name to the new name.
6330 * journald's SplitMode= now defaults to "uid". This makes sure
6331 that unprivileged users can access their own coredumps with
6332 coredumpctl without restrictions.
6334 * New kernel command line options "systemd.wants=" (for
6335 pulling an additional unit during boot), "systemd.mask="
6336 (for masking a specific unit for the boot), and
6337 "systemd.debug-shell" (for enabling the debug shell on tty9)
6338 have been added. This is implemented in the new generator
6339 "systemd-debug-generator".
6341 * systemd-nspawn will now by default filter a couple of
6342 syscalls for containers, among them those required for
6343 kernel module loading, direct x86 IO port access, swap
6344 management, and kexec. Most importantly though
6345 open_by_handle_at() is now prohibited for containers,
6346 closing a hole similar to a recently discussed vulnerability
6347 in docker regarding access to files on file hierarchies the
6348 container should normally not have access to. Note that, for
6349 nspawn, we generally make no security claims anyway (and
6350 this is explicitly documented in the man page), so this is
6351 just a fix for one of the most obvious problems.
6353 * A new man page file-hierarchy(7) has been added that
6354 contains a minimized, modernized version of the file system
6355 layout systemd expects, similar in style to the FHS
6356 specification or hier(5). A new tool systemd-path(1) has
6357 been added to query many of these paths for the local
6360 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
6361 longer done. Since the directory now has a per-user size
6362 limit, and is cleaned on logout this appears unnecessary,
6363 in particular since this now brings the lifecycle of this
6364 directory closer in line with how IPC objects are handled.
6366 * systemd.pc now exports a number of additional directories,
6367 including $libdir (which is useful to identify the library
6368 path for the primary architecture of the system), and a
6369 couple of drop-in directories.
6371 * udev's predictable network interface names now use the dev_port
6372 sysfs attribute, introduced in linux 3.15 instead of dev_id to
6373 distinguish between ports of the same PCI function. dev_id should
6374 only be used for ports using the same HW address, hence the need
6377 * machined has been updated to export the OS version of a
6378 container (read from /etc/os-release and
6379 /usr/lib/os-release) on the bus. This is now shown in
6380 "machinectl status" for a machine.
6382 * A new service setting RestartForceExitStatus= has been
6383 added. If configured to a set of exit signals or process
6384 return values, the service will be restarted when the main
6385 daemon process exits with any of them, regardless of the
6388 * systemctl's -H switch for connecting to remote systemd
6389 machines has been extended so that it may be used to
6390 directly connect to a specific container on the
6391 host. "systemctl -H root@foobar:waldi" will now connect as
6392 user "root" to host "foobar", and then proceed directly to
6393 the container named "waldi". Note that currently you have to
6394 authenticate as user "root" for this to work, as entering
6395 containers is a privileged operation.
6397 Contributions from: Andreas Henriksson, Benjamin Steinwender,
6398 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
6399 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
6400 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
6401 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
6402 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
6403 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
6404 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
6405 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
6406 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
6407 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
6408 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
6410 — Berlin, 2014-07-03
6414 * As an experimental feature, udev now tries to lock the
6415 disk device node (flock(LOCK_SH|LOCK_NB)) while it
6416 executes events for the disk or any of its partitions.
6417 Applications like partitioning programs can lock the
6418 disk device node (flock(LOCK_EX)) and claim temporary
6419 device ownership that way; udev will entirely skip all event
6420 handling for this disk and its partitions. If the disk
6421 was opened for writing, the close will trigger a partition
6422 table rescan in udev's "watch" facility, and if needed
6423 synthesize "change" events for the disk and all its partitions.
6424 This is now unconditionally enabled, and if it turns out to
6425 cause major problems, we might turn it on only for specific
6426 devices, or might need to disable it entirely. Device Mapper
6427 devices are excluded from this logic.
6429 * We temporarily dropped the "-l" switch for fsck invocations,
6430 since they collide with the flock() logic above. util-linux
6431 upstream has been changed already to avoid this conflict,
6432 and we will re-add "-l" as soon as util-linux with this
6433 change has been released.
6435 * The dependency on libattr has been removed. Since a long
6436 time, the extended attribute calls have moved to glibc, and
6437 libattr is thus unnecessary.
6439 * Virtualization detection works without privileges now. This
6440 means the systemd-detect-virt binary no longer requires
6441 CAP_SYS_PTRACE file capabilities, and our daemons can run
6442 with fewer privileges.
6444 * systemd-networkd now runs under its own "systemd-network"
6445 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
6446 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
6447 loses the ability to write to files owned by root this way.
6449 * Similarly, systemd-resolved now runs under its own
6450 "systemd-resolve" user with no capabilities remaining.
6452 * Similarly, systemd-bus-proxyd now runs under its own
6453 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
6455 * systemd-networkd gained support for setting up "veth"
6456 virtual Ethernet devices for container connectivity, as well
6457 as GRE and VTI tunnels.
6459 * systemd-networkd will no longer automatically attempt to
6460 manually load kernel modules necessary for certain tunnel
6461 transports. Instead, it is assumed the kernel loads them
6462 automatically when required. This only works correctly on
6463 very new kernels. On older kernels, please consider adding
6464 the kernel modules to /etc/modules-load.d/ as a work-around.
6466 * The resolv.conf file systemd-resolved generates has been
6467 moved to /run/systemd/resolve/. If you have a symlink from
6468 /etc/resolv.conf, it might be necessary to correct it.
6470 * Two new service settings, ProtectHome= and ProtectSystem=,
6471 have been added. When enabled, they will make the user data
6472 (such as /home) inaccessible or read-only and the system
6473 (such as /usr) read-only, for specific services. This allows
6474 very light-weight per-service sandboxing to avoid
6475 modifications of user data or system files from
6476 services. These two new switches have been enabled for all
6477 of systemd's long-running services, where appropriate.
6479 * Socket units gained new SocketUser= and SocketGroup=
6480 settings to set the owner user and group of AF_UNIX sockets
6481 and FIFOs in the file system.
6483 * Socket units gained a new RemoveOnStop= setting. If enabled,
6484 all FIFOS and sockets in the file system will be removed
6485 when the specific socket unit is stopped.
6487 * Socket units gained a new Symlinks= setting. It takes a list
6488 of symlinks to create to file system sockets or FIFOs
6489 created by the specific Unix sockets. This is useful to
6490 manage symlinks to socket nodes with the same lifecycle as
6493 * The /dev/log socket and /dev/initctl FIFO have been moved to
6494 /run, and have been replaced by symlinks. This allows
6495 connecting to these facilities even if PrivateDevices=yes is
6496 used for a service (which makes /dev/log itself unavailable,
6497 but /run is left). This also has the benefit of ensuring
6498 that /dev only contains device nodes, directories and
6499 symlinks, and nothing else.
6501 * sd-daemon gained two new calls sd_pid_notify() and
6502 sd_pid_notifyf(). They are similar to sd_notify() and
6503 sd_notifyf(), but allow overriding of the source PID of
6504 notification messages if permissions permit this. This is
6505 useful to send notify messages on behalf of a different
6506 process (for example, the parent process). The
6507 systemd-notify tool has been updated to make use of this
6508 when sending messages (so that notification messages now
6509 originate from the shell script invoking systemd-notify and
6510 not the systemd-notify process itself. This should minimize
6511 a race where systemd fails to associate notification
6512 messages to services when the originating process already
6515 * A new "on-abnormal" setting for Restart= has been added. If
6516 set, it will result in automatic restarts on all "abnormal"
6517 reasons for a process to exit, which includes unclean
6518 signals, core dumps, timeouts and watchdog timeouts, but
6519 does not include clean and unclean exit codes or clean
6520 signals. Restart=on-abnormal is an alternative for
6521 Restart=on-failure for services that shall be able to
6522 terminate and avoid restarts on certain errors, by
6523 indicating so with an unclean exit code. Restart=on-failure
6524 or Restart=on-abnormal is now the recommended setting for
6525 all long-running services.
6527 * If the InaccessibleDirectories= service setting points to a
6528 mount point (or if there are any submounts contained within
6529 it), it is now attempted to completely unmount it, to make
6530 the file systems truly unavailable for the respective
6533 * The ReadOnlyDirectories= service setting and
6534 systemd-nspawn's --read-only parameter are now recursively
6535 applied to all submounts, too.
6537 * Mount units may now be created transiently via the bus APIs.
6539 * The support for SysV and LSB init scripts has been removed
6540 from the systemd daemon itself. Instead, it is now
6541 implemented as a generator that creates native systemd units
6542 from these scripts when needed. This enables us to remove a
6543 substantial amount of legacy code from PID 1, following the
6544 fact that many distributions only ship a very small number
6545 of LSB/SysV init scripts nowadays.
6547 * Privileged Xen (dom0) domains are not considered
6548 virtualization anymore by the virtualization detection
6549 logic. After all, they generally have unrestricted access to
6550 the hardware and usually are used to manage the unprivileged
6553 * systemd-tmpfiles gained a new "C" line type, for copying
6554 files or entire directories.
6556 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
6557 lines. So far, they have been non-globbing versions of the
6558 latter, and have thus been redundant. In future, it is
6559 recommended to only use "z". "m" has hence been removed
6560 from the documentation, even though it stays supported.
6562 * A tmpfiles snippet to recreate the most basic structure in
6563 /var has been added. This is enough to create the /var/run →
6564 /run symlink and create a couple of structural
6565 directories. This allows systems to boot up with an empty or
6566 volatile /var. Of course, while with this change, the core OS
6567 now is capable with dealing with a volatile /var, not all
6568 user services are ready for it. However, we hope that sooner
6569 or later, many service daemons will be changed upstream so
6570 that they are able to automatically create their necessary
6571 directories in /var at boot, should they be missing. This is
6572 the first step to allow state-less systems that only require
6573 the vendor image for /usr to boot.
6575 * systemd-nspawn has gained a new --tmpfs= switch to mount an
6576 empty tmpfs instance to a specific directory. This is
6577 particularly useful for making use of the automatic
6578 reconstruction of /var (see above), by passing --tmpfs=/var.
6580 * Access modes specified in tmpfiles snippets may now be
6581 prefixed with "~", which indicates that they shall be masked
6582 by whether the existing file or directory is currently
6583 writable, readable or executable at all. Also, if specified,
6584 the sgid/suid/sticky bits will be masked for all
6587 * A new passive target unit "network-pre.target" has been
6588 added which is useful for services that shall run before any
6589 network is configured, for example firewall scripts.
6591 * The "floppy" group that previously owned the /dev/fd*
6592 devices is no longer used. The "disk" group is now used
6593 instead. Distributions should probably deprecate usage of
6596 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
6597 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
6598 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
6599 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
6600 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
6601 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
6604 — Berlin, 2014-06-11
6608 * A new "systemd-timesyncd" daemon has been added for
6609 synchronizing the system clock across the network. It
6610 implements an SNTP client. In contrast to NTP
6611 implementations such as chrony or the NTP reference server,
6612 this only implements a client side, and does not bother with
6613 the full NTP complexity, focusing only on querying time from
6614 one remote server and synchronizing the local clock to
6615 it. Unless you intend to serve NTP to networked clients or
6616 want to connect to local hardware clocks, this simple NTP
6617 client should be more than appropriate for most
6618 installations. The daemon runs with minimal privileges, and
6619 has been hooked up with networkd to only operate when
6620 network connectivity is available. The daemon saves the
6621 current clock to disk every time a new NTP sync has been
6622 acquired, and uses this to possibly correct the system clock
6623 early at bootup, in order to accommodate for systems that
6624 lack an RTC such as the Raspberry Pi and embedded devices,
6625 and to make sure that time monotonically progresses on these
6626 systems, even if it is not always correct. To make use of
6627 this daemon, a new system user and group "systemd-timesync"
6628 needs to be created on installation of systemd.
6630 * The queue "seqnum" interface of libudev has been disabled, as
6631 it was generally incompatible with device namespacing as
6632 sequence numbers of devices go "missing" if the devices are
6633 part of a different namespace.
6635 * "systemctl list-timers" and "systemctl list-sockets" gained
6636 a --recursive switch for showing units of these types also
6637 for all local containers, similar in style to the already
6638 supported --recursive switch for "systemctl list-units".
6640 * A new RebootArgument= setting has been added for service
6641 units, which may be used to specify a kernel reboot argument
6642 to use when triggering reboots with StartLimitAction=.
6644 * A new FailureAction= setting has been added for service
6645 units which may be used to specify an operation to trigger
6646 when a service fails. This works similarly to
6647 StartLimitAction=, but unlike it, controls what is done
6648 immediately rather than only after several attempts to
6649 restart the service in question.
6651 * hostnamed got updated to also expose the kernel name,
6652 release, and version on the bus. This is useful for
6653 executing commands like hostnamectl with the -H switch.
6654 systemd-analyze makes use of this to properly display
6655 details when running non-locally.
6657 * The bootchart tool can now show cgroup information in the
6658 graphs it generates.
6660 * The CFS CPU quota cgroup attribute is now exposed for
6661 services. The new CPUQuota= switch has been added for this
6662 which takes a percentage value. Setting this will have the
6663 result that a service may never get more CPU time than the
6664 specified percentage, even if the machine is otherwise idle.
6666 * systemd-networkd learned IPIP and SIT tunnel support.
6668 * LSB init scripts exposing a dependency on $network will now
6669 get a dependency on network-online.target rather than simply
6670 network.target. This should bring LSB handling closer to
6671 what it was on SysV systems.
6673 * A new fsck.repair= kernel option has been added to control
6674 how fsck shall deal with unclean file systems at boot.
6676 * The (.ini) configuration file parser will now silently
6677 ignore sections whose name begins with "X-". This may be
6678 used to maintain application-specific extension sections in unit
6681 * machined gained a new API to query the IP addresses of
6682 registered containers. "machinectl status" has been updated
6683 to show these addresses in its output.
6685 * A new call sd_uid_get_display() has been added to the
6686 sd-login APIs for querying the "primary" session of a
6687 user. The "primary" session of the user is elected from the
6688 user's sessions and generally a graphical session is
6689 preferred over a text one.
6691 * A minimal systemd-resolved daemon has been added. It
6692 currently simply acts as a companion to systemd-networkd and
6693 manages resolv.conf based on per-interface DNS
6694 configuration, possibly supplied via DHCP. In the long run
6695 we hope to extend this into a local DNSSEC enabled DNS and
6698 * The systemd-networkd-wait-online tool is now enabled by
6699 default. It will delay network-online.target until a network
6700 connection has been configured. The tool primarily integrates
6701 with networkd, but will also make a best effort to make sense
6702 of network configuration performed in some other way.
6704 * Two new service options StartupCPUShares= and
6705 StartupBlockIOWeight= have been added that work similarly to
6706 CPUShares= and BlockIOWeight= however only apply during
6707 system startup. This is useful to prioritize certain services
6708 differently during bootup than during normal runtime.
6710 * hostnamed has been changed to prefer the statically
6711 configured hostname in /etc/hostname (unless set to
6712 'localhost' or empty) over any dynamic one supplied by
6713 dhcp. With this change, the rules for picking the hostname
6714 match more closely the rules of other configuration settings
6715 where the local administrator's configuration in /etc always
6716 overrides any other settings.
6718 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
6719 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
6720 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
6721 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
6722 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
6723 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
6724 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
6725 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
6726 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
6727 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
6728 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
6729 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
6730 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
6731 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
6732 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
6733 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6736 — Beijing, 2014-05-28
6740 * When restoring the screen brightness at boot, stay away from
6741 the darkest setting or from the lowest 5% of the available
6742 range, depending on which is the larger value of both. This
6743 should effectively protect the user from rebooting into a
6744 black screen, should the brightness have been set to minimum
6747 * sd-login gained a new sd_machine_get_class() call to
6748 determine the class ("vm" or "container") of a machine
6749 registered with machined.
6751 * sd-login gained new calls
6752 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
6753 to query the identity of the peer of a local AF_UNIX
6754 connection. They operate similarly to their sd_pid_get_xyz()
6757 * PID 1 will now maintain a system-wide system state engine
6758 with the states "starting", "running", "degraded",
6759 "maintenance", "stopping". These states are bound to system
6760 startup, normal runtime, runtime with at least one failed
6761 service, rescue/emergency mode and system shutdown. This
6762 state is shown in the "systemctl status" output when no unit
6763 name is passed. It is useful to determine system state, in
6764 particularly when doing so for many systems or containers at
6767 * A new command "list-machines" has been added to "systemctl"
6768 that lists all local OS containers and shows their system
6769 state (see above), if systemd runs inside of them.
6771 * systemctl gained a new "-r" switch to recursively enumerate
6772 units on all local containers, when used with the
6773 "list-unit" command (which is the default one that is
6774 executed when no parameters are specified).
6776 * The GPT automatic partition discovery logic will now honour
6777 two GPT partition flags: one may be set on a partition to
6778 cause it to be mounted read-only, and the other may be set
6779 on a partition to ignore it during automatic discovery.
6781 * Two new GPT type UUIDs have been added for automatic root
6782 partition discovery, for 32-bit and 64-bit ARM. This is not
6783 particularly useful for discovering the root directory on
6784 these architectures during bare-metal boots (since UEFI is
6785 not common there), but still very useful to allow booting of
6786 ARM disk images in nspawn with the -i option.
6788 * MAC addresses of interfaces created with nspawn's
6789 --network-interface= switch will now be generated from the
6790 machine name, and thus be stable between multiple invocations
6793 * logind will now automatically remove all IPC objects owned
6794 by a user if she or he fully logs out. This makes sure that
6795 users who are logged out cannot continue to consume IPC
6796 resources. This covers SysV memory, semaphores and message
6797 queues as well as POSIX shared memory and message
6798 queues. Traditionally, SysV and POSIX IPC had no lifecycle
6799 limits. With this functionality, that is corrected. This may
6800 be turned off by using the RemoveIPC= switch of logind.conf.
6802 * The systemd-machine-id-setup and tmpfiles tools gained a
6803 --root= switch to operate on a specific root directory,
6806 * journald can now forward logged messages to the TTYs of all
6807 logged in users ("wall"). This is the default for all
6808 emergency messages now.
6810 * A new tool systemd-journal-remote has been added to stream
6811 journal log messages across the network.
6813 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
6814 controller trees are mounted into it. Note that the
6815 directories mounted beneath it are not read-only. This is a
6816 security measure and is particularly useful because glibc
6817 actually includes a search logic to pick any tmpfs it can
6818 find to implement shm_open() if /dev/shm is not available
6819 (which it might very well be in namespaced setups).
6821 * machinectl gained a new "poweroff" command to cleanly power
6822 down a local OS container.
6824 * The PrivateDevices= unit file setting will now also drop the
6825 CAP_MKNOD capability from the capability bound set, and
6826 imply DevicePolicy=closed.
6828 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
6829 comprehensively on all long-running systemd services where
6830 this is appropriate.
6832 * systemd-udevd will now run in a disassociated mount
6833 namespace. To mount directories from udev rules, make sure to
6834 pull in mount units via SYSTEMD_WANTS properties.
6836 * The kdbus support gained support for uploading policy into
6837 the kernel. sd-bus gained support for creating "monitoring"
6838 connections that can eavesdrop into all bus communication
6839 for debugging purposes.
6841 * Timestamps may now be specified in seconds since the UNIX
6842 epoch Jan 1st, 1970 by specifying "@" followed by the value
6845 * Native tcpwrap support in systemd has been removed. tcpwrap
6846 is old code, not really maintained anymore and has serious
6847 shortcomings, and better options such as firewalls
6848 exist. For setups that require tcpwrap usage, please
6849 consider invoking your socket-activated service via tcpd,
6850 like on traditional inetd.
6852 * A new system.conf configuration option
6853 DefaultTimerAccuracySec= has been added that controls the
6854 default AccuracySec= setting of .timer units.
6856 * Timer units gained a new WakeSystem= switch. If enabled,
6857 timers configured this way will cause the system to resume
6858 from system suspend (if the system supports that, which most
6861 * Timer units gained a new Persistent= switch. If enabled,
6862 timers configured this way will save to disk when they have
6863 been last triggered. This information is then used on next
6864 reboot to possible execute overdue timer events, that
6865 could not take place because the system was powered off.
6866 This enables simple anacron-like behaviour for timer units.
6868 * systemctl's "list-timers" will now also list the time a
6869 timer unit was last triggered in addition to the next time
6870 it will be triggered.
6872 * systemd-networkd will now assign predictable IPv4LL
6873 addresses to its local interfaces.
6875 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
6876 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
6877 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
6878 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
6879 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
6880 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
6881 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
6882 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
6885 — Berlin, 2014-03-25
6889 * A new unit file setting RestrictAddressFamilies= has been
6890 added to restrict which socket address families unit
6891 processes gain access to. This takes address family names
6892 like "AF_INET" or "AF_UNIX", and is useful to minimize the
6893 attack surface of services via exotic protocol stacks. This
6894 is built on seccomp system call filters.
6896 * Two new unit file settings RuntimeDirectory= and
6897 RuntimeDirectoryMode= have been added that may be used to
6898 manage a per-daemon runtime directories below /run. This is
6899 an alternative for setting up directory permissions with
6900 tmpfiles snippets, and has the advantage that the runtime
6901 directory's lifetime is bound to the daemon runtime and that
6902 the daemon starts up with an empty directory each time. This
6903 is particularly useful when writing services that drop
6904 privileges using the User= or Group= setting.
6906 * The DeviceAllow= unit setting now supports globbing for
6907 matching against device group names.
6909 * The systemd configuration file system.conf gained new
6910 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
6911 DefaultMemoryAccounting= to globally turn on/off accounting
6912 for specific resources (cgroups) for all units. These
6913 settings may still be overridden individually in each unit
6916 * systemd-gpt-auto-generator is now able to discover /srv and
6917 root partitions in addition to /home and swap partitions. It
6918 also supports LUKS-encrypted partitions now. With this in
6919 place, automatic discovery of partitions to mount following
6920 the Discoverable Partitions Specification
6921 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
6922 is now a lot more complete. This allows booting without
6923 /etc/fstab and without root= on the kernel command line on
6924 systems prepared appropriately.
6926 * systemd-nspawn gained a new --image= switch which allows
6927 booting up disk images and Linux installations on any block
6928 device that follow the Discoverable Partitions Specification
6929 (see above). This means that installations made with
6930 appropriately updated installers may now be started and
6931 deployed using container managers, completely
6932 unmodified. (We hope that libvirt-lxc will add support for
6933 this feature soon, too.)
6935 * systemd-nspawn gained a new --network-macvlan= setting to
6936 set up a private macvlan interface for the
6937 container. Similarly, systemd-networkd gained a new
6938 Kind=macvlan setting in .netdev files.
6940 * systemd-networkd now supports configuring local addresses
6943 * A new tool systemd-network-wait-online has been added to
6944 synchronously wait for network connectivity using
6947 * The sd-bus.h bus API gained a new sd_bus_track object for
6948 tracking the lifecycle of bus peers. Note that sd-bus.h is
6949 still not a public API though (unless you specify
6950 --enable-kdbus on the configure command line, which however
6951 voids your warranty and you get no API stability guarantee).
6953 * The $XDG_RUNTIME_DIR runtime directories for each user are
6954 now individual tmpfs instances, which has the benefit of
6955 introducing separate pools for each user, with individual
6956 size limits, and thus making sure that unprivileged clients
6957 can no longer negatively impact the system or other users by
6958 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
6959 RuntimeDirectorySize= has been introduced that allows
6960 controlling the default size limit for all users. It
6961 defaults to 10% of the available physical memory. This is no
6962 replacement for quotas on tmpfs though (which the kernel
6963 still does not support), as /dev/shm and /tmp are still
6964 shared resources used by both the system and unprivileged
6967 * logind will now automatically turn off automatic suspending
6968 on laptop lid close when more than one display is
6969 connected. This was previously expected to be implemented
6970 individually in desktop environments (such as GNOME),
6971 however has been added to logind now, in order to fix a
6972 boot-time race where a desktop environment might not have
6973 been started yet and thus not been able to take an inhibitor
6974 lock at the time where logind already suspends the system
6975 due to a closed lid.
6977 * logind will now wait at least 30s after each system
6978 suspend/resume cycle, and 3min after system boot before
6979 suspending the system due to a closed laptop lid. This
6980 should give USB docking stations and similar enough time to
6981 be probed and configured after system resume and boot in
6982 order to then act as suspend blocker.
6984 * systemd-run gained a new --property= setting which allows
6985 initialization of resource control properties (and others)
6986 for the created scope or service unit. Example: "systemd-run
6987 --property=BlockIOWeight=10 updatedb" may be used to run
6988 updatedb at a low block IO scheduling weight.
6990 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
6991 now also work in --scope mode.
6993 * When systemd is compiled with kdbus support, basic support
6994 for enforced policies is now in place. (Note that enabling
6995 kdbus still voids your warranty and no API compatibility
6998 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
6999 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7000 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
7001 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
7002 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
7003 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
7004 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
7005 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
7006 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
7007 Zbigniew Jędrzejewski-Szmek
7009 — Berlin, 2014-03-12
7013 * systemd will now relabel /dev after loading the SMACK policy
7014 according to SMACK rules.
7016 * A new unit file option AppArmorProfile= has been added to
7017 set the AppArmor profile for the processes of a unit.
7019 * A new condition check ConditionArchitecture= has been added
7020 to conditionalize units based on the system architecture, as
7021 reported by uname()'s "machine" field.
7023 * systemd-networkd now supports matching on the system
7024 virtualization, architecture, kernel command line, host name
7027 * logind is now a lot more aggressive when suspending the
7028 machine due to a closed laptop lid. Instead of acting only
7029 on the lid close action, it will continuously watch the lid
7030 status and act on it. This is useful for laptops where the
7031 power button is on the outside of the chassis so that it can
7032 be reached without opening the lid (such as the Lenovo
7033 Yoga). On those machines, logind will now immediately
7034 re-suspend the machine if the power button has been
7035 accidentally pressed while the laptop was suspended and in a
7036 backpack or similar.
7038 * logind will now watch SW_DOCK switches and inhibit reaction
7039 to the lid switch if it is pressed. This means that logind
7040 will not suspend the machine anymore if the lid is closed
7041 and the system is docked, if the laptop supports SW_DOCK
7042 notifications via the input layer. Note that ACPI docking
7043 stations do not generate this currently. Also note that this
7044 logic is usually not fully sufficient and Desktop
7045 Environments should take a lid switch inhibitor lock when an
7046 external display is connected, as systemd will not watch
7049 * nspawn will now make use of the devices cgroup controller by
7050 default, and only permit creation of and access to the usual
7051 API device nodes like /dev/null or /dev/random, as well as
7052 access to (but not creation of) the pty devices.
7054 * We will now ship a default .network file for
7055 systemd-networkd that automatically configures DHCP for
7056 network interfaces created by nspawn's --network-veth or
7057 --network-bridge= switches.
7059 * systemd will now understand the usual M, K, G, T suffixes
7060 according to SI conventions (i.e. to the base 1000) when
7061 referring to throughput and hardware metrics. It will stay
7062 with IEC conventions (i.e. to the base 1024) for software
7063 metrics, according to what is customary according to
7064 Wikipedia. We explicitly document which base applies for
7065 each configuration option.
7067 * The DeviceAllow= setting in unit files now supports a syntax
7068 to whitelist an entire group of devices node majors at once,
7069 based on the /proc/devices listing. For example, with the
7070 string "char-pts", it is now possible to whitelist all
7071 current and future pseudo-TTYs at once.
7073 * sd-event learned a new "post" event source. Event sources of
7074 this type are triggered by the dispatching of any event
7075 source of a type that is not "post". This is useful for
7076 implementing clean-up and check event sources that are
7077 triggered by other work being done in the program.
7079 * systemd-networkd is no longer statically enabled, but uses
7080 the usual [Install] sections so that it can be
7081 enabled/disabled using systemctl. It still is enabled by
7084 * When creating a veth interface pair with systemd-nspawn, the
7085 host side will now be prefixed with "vb-" if
7086 --network-bridge= is used, and with "ve-" if --network-veth
7087 is used. This way, it is easy to distinguish these cases on
7088 the host, for example to apply different configuration to
7089 them with systemd-networkd.
7091 * The compatibility libraries for libsystemd-journal.so,
7092 libsystem-id128.so, libsystemd-login.so and
7093 libsystemd-daemon.so do not make use of IFUNC
7094 anymore. Instead, we now build libsystemd.so multiple times
7095 under these alternative names. This means that the footprint
7096 is drastically increased, but given that these are
7097 transitional compatibility libraries, this should not matter
7098 much. This change has been made necessary to support the ARM
7099 platform for these compatibility libraries, as the ARM
7100 toolchain is not really at the same level as the toolchain
7101 for other architectures like x86 and does not support
7102 IFUNC. Please make sure to use --enable-compat-libs only
7103 during a transitional period!
7105 * The .include syntax has been deprecated and is not documented
7106 anymore. Drop-in files in .d directories should be used instead.
7108 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
7109 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7110 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
7111 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
7112 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
7113 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
7114 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
7115 Zbigniew Jędrzejewski-Szmek
7117 — Berlin, 2014-02-24
7121 * A new component "systemd-networkd" has been added that can
7122 be used to configure local network interfaces statically or
7123 via DHCP. It is capable of bringing up bridges, VLANs, and
7124 bonding. Currently, no hook-ups for interactive network
7125 configuration are provided. Use this for your initrd,
7126 container, embedded, or server setup if you need a simple,
7127 yet powerful, network configuration solution. This
7128 configuration subsystem is quite nifty, as it allows wildcard
7129 hotplug matching in interfaces. For example, with a single
7130 configuration snippet, you can configure that all Ethernet
7131 interfaces showing up are automatically added to a bridge,
7132 or similar. It supports link-sensing and more.
7134 * A new tool "systemd-socket-proxyd" has been added which can
7135 act as a bidirectional proxy for TCP sockets. This is
7136 useful for adding socket activation support to services that
7137 do not actually support socket activation, including virtual
7138 machines and the like.
7140 * Add a new tool to save/restore rfkill state on
7143 * Save/restore state of keyboard backlights in addition to
7144 display backlights on shutdown/boot.
7146 * udev learned a new SECLABEL{} construct to label device
7147 nodes with a specific security label when they appear. For
7148 now, only SECLABEL{selinux} is supported, but the syntax is
7149 prepared for additional security frameworks.
7151 * udev gained a new scheme to configure link-level attributes
7152 from files in /etc/systemd/network/*.link. These files can
7153 match against MAC address, device path, driver name and type,
7154 and will apply attributes like the naming policy, link speed,
7155 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
7156 address assignment policy (randomized, ...).
7158 * The configuration of network interface naming rules for
7159 "permanent interface names" has changed: a new NamePolicy=
7160 setting in the [Link] section of .link files determines the
7161 priority of possible naming schemes (onboard, slot, MAC,
7162 path). The default value of this setting is determined by
7163 /usr/lib/net/links/99-default.link. Old
7164 80-net-name-slot.rules udev configuration file has been
7165 removed, so local configuration overriding this file should
7166 be adapted to override 99-default.link instead.
7168 * When the User= switch is used in a unit file, also
7169 initialize $SHELL= based on the user database entry.
7171 * systemd no longer depends on libdbus. All communication is
7172 now done with sd-bus, systemd's low-level bus library
7175 * kdbus support has been added to PID 1 itself. When kdbus is
7176 enabled, this causes PID 1 to set up the system bus and
7177 enable support for a new ".busname" unit type that
7178 encapsulates bus name activation on kdbus. It works a little
7179 bit like ".socket" units, except for bus names. A new
7180 generator has been added that converts classic dbus1 service
7181 activation files automatically into native systemd .busname
7184 * sd-bus: add a light-weight vtable implementation that allows
7185 defining objects on the bus with a simple static const
7186 vtable array of its methods, signals and properties.
7188 * systemd will not generate or install static dbus
7189 introspection data anymore to /usr/share/dbus-1/interfaces,
7190 as the precise format of these files is unclear, and
7191 nothing makes use of it.
7193 * A proxy daemon is now provided to proxy clients connecting
7194 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
7195 compatibility with classic D-Bus.
7197 * A bus driver implementation has been added that supports the
7198 classic D-Bus bus driver calls on kdbus, also for
7199 compatibility purposes.
7201 * A new API "sd-event.h" has been added that implements a
7202 minimal event loop API built around epoll. It provides a
7203 couple of features that direct epoll usage is lacking:
7204 prioritization of events, scales to large numbers of timer
7205 events, per-event timer slack (accuracy), system-wide
7206 coalescing of timer events, exit handlers, watchdog
7207 supervision support using systemd's sd_notify() API, child
7210 * A new API "sd-rntl.h" has been added that provides an API
7211 around the route netlink interface of the kernel, similar in
7212 style to "sd-bus.h".
7214 * A new API "sd-dhcp-client.h" has been added that provides a
7215 small DHCPv4 client-side implementation. This is used by
7218 * There is a new kernel command line option
7219 "systemd.restore_state=0|1". When set to "0", none of the
7220 systemd tools will restore saved runtime state to hardware
7221 devices. More specifically, the rfkill and backlight states
7224 * The FsckPassNo= compatibility option in mount/service units
7225 has been removed. The fstab generator will now add the
7226 necessary dependencies automatically, and does not require
7227 PID1's support for that anymore.
7229 * journalctl gained a new switch, --list-boots, that lists
7230 recent boots with their times and boot IDs.
7232 * The various tools like systemctl, loginctl, timedatectl,
7233 busctl, systemd-run, ... have gained a new switch "-M" to
7234 connect to a specific, local OS container (as direct
7235 connection, without requiring SSH). This works on any
7236 container that is registered with machined, such as those
7237 created by libvirt-lxc or nspawn.
7239 * systemd-run and systemd-analyze also gained support for "-H"
7240 to connect to remote hosts via SSH. This is particularly
7241 useful for systemd-run because it enables queuing of jobs
7242 onto remote systems.
7244 * machinectl gained a new command "login" to open a getty
7245 login in any local container. This works with any container
7246 that is registered with machined (such as those created by
7247 libvirt-lxc or nspawn), and which runs systemd inside.
7249 * machinectl gained a new "reboot" command that may be used to
7250 trigger a reboot on a specific container that is registered
7251 with machined. This works on any container that runs an init
7252 system of some kind.
7254 * systemctl gained a new "list-timers" command to print a nice
7255 listing of installed timer units with the times they elapse
7258 * Alternative reboot() parameters may now be specified on the
7259 "systemctl reboot" command line and are passed to the
7260 reboot() system call.
7262 * systemctl gained a new --job-mode= switch to configure the
7263 mode to queue a job with. This is a more generic version of
7264 --fail, --irreversible, and --ignore-dependencies, which are
7265 still available but not advertised anymore.
7267 * /etc/systemd/system.conf gained new settings to configure
7268 various default timeouts of units, as well as the default
7269 start limit interval and burst. These may still be overridden
7272 * PID1 will now export on the bus profile data of the security
7273 policy upload process (such as the SELinux policy upload to
7276 * journald: when forwarding logs to the console, include
7277 timestamps (following the setting in
7278 /sys/module/printk/parameters/time).
7280 * OnCalendar= in timer units now understands the special
7281 strings "yearly" and "annually". (Both are equivalent)
7283 * The accuracy of timer units is now configurable with the new
7284 AccuracySec= setting. It defaults to 1min.
7286 * A new dependency type JoinsNamespaceOf= has been added that
7287 allows running two services within the same /tmp and network
7288 namespace, if PrivateNetwork= or PrivateTmp= are used.
7290 * A new command "cat" has been added to systemctl. It outputs
7291 the original unit file of a unit, and concatenates the
7292 contents of additional "drop-in" unit file snippets, so that
7293 the full configuration is shown.
7295 * systemctl now supports globbing on the various "list-xyz"
7296 commands, like "list-units" or "list-sockets", as well as on
7297 those commands which take multiple unit names.
7299 * journalctl's --unit= switch gained support for globbing.
7301 * All systemd daemons now make use of the watchdog logic so
7302 that systemd automatically notices when they hang.
7304 * If the $container_ttys environment variable is set,
7305 getty-generator will automatically spawn a getty for each
7306 listed tty. This is useful for container managers to request
7307 login gettys to be spawned on as many ttys as needed.
7309 * %h, %s, %U specifier support is not available anymore when
7310 used in unit files for PID 1. This is because NSS calls are
7311 not safe from PID 1. They stay available for --user
7312 instances of systemd, and as special case for the root user.
7314 * loginctl gained a new "--no-legend" switch to turn off output
7317 * The "sd-login.h" API gained three new calls:
7318 sd_session_is_remote(), sd_session_get_remote_user(),
7319 sd_session_get_remote_host() to query information about
7322 * The udev hardware database now also carries vendor/product
7323 information of SDIO devices.
7325 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
7326 determine whether watchdog notifications are requested by
7329 * Socket-activated per-connection services now include a
7330 short description of the connection parameters in the
7333 * tmpfiles gained a new "--boot" option. When this is not used,
7334 only lines where the command character is not suffixed with
7335 "!" are executed. When this option is specified, those
7336 options are executed too. This partitions tmpfiles
7337 directives into those that can be safely executed at any
7338 time, and those which should be run only at boot (for
7339 example, a line that creates /run/nologin).
7341 * A new API "sd-resolve.h" has been added which provides a simple
7342 asynchronous wrapper around glibc NSS host name resolution
7343 calls, such as getaddrinfo(). In contrast to glibc's
7344 getaddrinfo_a(), it does not use signals. In contrast to most
7345 other asynchronous name resolution libraries, this one does
7346 not reimplement DNS, but reuses NSS, so that alternate
7347 host name resolution systems continue to work, such as mDNS,
7348 LDAP, etc. This API is based on libasyncns, but it has been
7349 cleaned up for inclusion in systemd.
7351 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
7352 "sd-daemon.h" are no longer found in individual libraries
7353 libsystemd-journal.so, libsystemd-login.so,
7354 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
7355 merged them into a single library, libsystemd.so, which
7356 provides all symbols. The reason for this is cyclic
7357 dependencies, as these libraries tend to use each other's
7358 symbols. So far, we have managed to workaround that by linking
7359 a copy of a good part of our code into each of these
7360 libraries again and again, which, however, makes certain
7361 things hard to do, like sharing static variables. Also, it
7362 substantially increases footprint. With this change, there
7363 is only one library for the basic APIs systemd
7364 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
7365 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
7366 library as well, however are subject to the --enable-kdbus
7367 switch (see below). Note that "sd-dhcp-client.h" is not part
7368 of this library (this is because it only consumes, never
7369 provides, services of/to other APIs). To make the transition
7370 easy from the separate libraries to the unified one, we
7371 provide the --enable-compat-libs compile-time switch which
7372 will generate stub libraries that are compatible with the
7373 old ones but redirect all calls to the new one.
7375 * All of the kdbus logic and the new APIs "sd-bus.h",
7376 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
7377 and "sd-utf8.h" are compile-time optional via the
7378 "--enable-kdbus" switch, and they are not compiled in by
7379 default. To make use of kdbus, you have to explicitly enable
7380 the switch. Note however, that neither the kernel nor the
7381 userspace API for all of this is considered stable yet. We
7382 want to maintain the freedom to still change the APIs for
7383 now. By specifying this build-time switch, you acknowledge
7384 that you are aware of the instability of the current
7387 * Also, note that while kdbus is pretty much complete,
7388 it lacks one thing: proper policy support. This means you
7389 can build a fully working system with all features; however,
7390 it will be highly insecure. Policy support will be added in
7391 one of the next releases, at the same time that we will
7392 declare the APIs stable.
7394 * When the kernel command line argument "kdbus" is specified,
7395 systemd will automatically load the kdbus.ko kernel module. At
7396 this stage of development, it is only useful for testing kdbus
7397 and should not be used in production. Note: if "--enable-kdbus"
7398 is specified, and the kdbus.ko kernel module is available, and
7399 "kdbus" is added to the kernel command line, the entire system
7400 runs with kdbus instead of dbus-daemon, with the above mentioned
7401 problem of missing the system policy enforcement. Also a future
7402 version of kdbus.ko or a newer systemd will not be compatible with
7403 each other, and will unlikely be able to boot the machine if only
7404 one of them is updated.
7406 * systemctl gained a new "import-environment" command which
7407 uploads the caller's environment (or parts thereof) into the
7408 service manager so that it is inherited by services started
7409 by the manager. This is useful to upload variables like
7410 $DISPLAY into the user service manager.
7412 * A new PrivateDevices= switch has been added to service units
7413 which allows running a service with a namespaced /dev
7414 directory that does not contain any device nodes for
7415 physical devices. More specifically, it only includes devices
7416 such as /dev/null, /dev/urandom, and /dev/zero which are API
7419 * logind has been extended to support behaviour like VT
7420 switching on seats that do not support a VT. This makes
7421 multi-session available on seats that are not the first seat
7422 (seat0), and on systems where kernel support for VTs has
7423 been disabled at compile-time.
7425 * If a process holds a delay lock for system sleep or shutdown
7426 and fails to release it in time, we will now log its
7427 identity. This makes it easier to identify processes that
7428 cause slow suspends or power-offs.
7430 * When parsing /etc/crypttab, support for a new key-slot=
7431 option as supported by Debian is added. It allows indicating
7432 which LUKS slot to use on disk, speeding up key loading.
7434 * The sd_journal_sendv() API call has been checked and
7435 officially declared to be async-signal-safe so that it may
7436 be invoked from signal handlers for logging purposes.
7438 * Boot-time status output is now enabled automatically after a
7439 short timeout if boot does not progress, in order to give
7440 the user an indication what she or he is waiting for.
7442 * The boot-time output has been improved to show how much time
7443 remains until jobs expire.
7445 * The KillMode= switch in service units gained a new possible
7446 value "mixed". If set, and the unit is shut down, then the
7447 initial SIGTERM signal is sent only to the main daemon
7448 process, while the following SIGKILL signal is sent to
7449 all remaining processes of the service.
7451 * When a scope unit is registered, a new property "Controller"
7452 may be set. If set to a valid bus name, systemd will send a
7453 RequestStop() signal to this name when it would like to shut
7454 down the scope. This may be used to hook manager logic into
7455 the shutdown logic of scope units. Also, scope units may now
7456 be put in a special "abandoned" state, in which case the
7457 manager process which created them takes no further
7458 responsibilities for it.
7460 * When reading unit files, systemd will now verify
7461 the access mode of these files, and warn about certain
7462 suspicious combinations. This has been added to make it
7463 easier to track down packaging bugs where unit files are
7464 marked executable or world-writable.
7466 * systemd-nspawn gained a new "--setenv=" switch to set
7467 container-wide environment variables. The similar option in
7468 systemd-activate was renamed from "--environment=" to
7469 "--setenv=" for consistency.
7471 * systemd-nspawn has been updated to create a new kdbus domain
7472 for each container that is invoked, thus allowing each
7473 container to have its own set of system and user buses,
7474 independent of the host.
7476 * systemd-nspawn gained a new --drop-capability= switch to run
7477 the container with less capabilities than the default. Both
7478 --drop-capability= and --capability= now take the special
7479 string "all" for dropping or keeping all capabilities.
7481 * systemd-nspawn gained new switches for executing containers
7482 with specific SELinux labels set.
7484 * systemd-nspawn gained a new --quiet switch to not generate
7485 any additional output but the container's own console
7488 * systemd-nspawn gained a new --share-system switch to run a
7489 container without PID namespacing enabled.
7491 * systemd-nspawn gained a new --register= switch to control
7492 whether the container is registered with systemd-machined or
7493 not. This is useful for containers that do not run full
7494 OS images, but only specific apps.
7496 * systemd-nspawn gained a new --keep-unit which may be used
7497 when invoked as the only program from a service unit, and
7498 results in registration of the unit service itself in
7499 systemd-machined, instead of a newly opened scope unit.
7501 * systemd-nspawn gained a new --network-interface= switch for
7502 moving arbitrary interfaces to the container. The new
7503 --network-veth switch creates a virtual Ethernet connection
7504 between host and container. The new --network-bridge=
7505 switch then allows assigning the host side of this virtual
7506 Ethernet connection to a bridge device.
7508 * systemd-nspawn gained a new --personality= switch for
7509 setting the kernel personality for the container. This is
7510 useful when running a 32-bit container on a 64-bit host. A
7511 similar option Personality= is now also available for service
7514 * logind will now also track a "Desktop" identifier for each
7515 session which encodes the desktop environment of it. This is
7516 useful for desktop environments that want to identify
7517 multiple running sessions of itself easily.
7519 * A new SELinuxContext= setting for service units has been
7520 added that allows setting a specific SELinux execution
7521 context for a service.
7523 * Most systemd client tools will now honour $SYSTEMD_LESS for
7524 settings of the "less" pager. By default, these tools will
7525 override $LESS to allow certain operations to work, such as
7526 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
7527 influence this logic.
7529 * systemd's "seccomp" hook-up has been changed to make use of
7530 the libseccomp library instead of using its own
7531 implementation. This has benefits for portability among
7534 * For usage together with SystemCallFilter=, a new
7535 SystemCallErrorNumber= setting has been introduced that
7536 allows configuration of a system error number to be returned
7537 on filtered system calls, instead of immediately killing the
7538 process. Also, SystemCallArchitectures= has been added to
7539 limit access to system calls of a particular architecture
7540 (in order to turn off support for unused secondary
7541 architectures). There is also a global
7542 SystemCallArchitectures= setting in system.conf now to turn
7543 off support for non-native system calls system-wide.
7545 * systemd requires a kernel with a working name_to_handle_at(),
7546 please see the kernel config requirements in the README file.
7548 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
7549 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
7550 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
7551 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
7552 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
7553 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
7554 Elia Pinto, Florian Weimer, George McCollister, Goffredo
7555 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
7556 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
7557 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
7558 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
7559 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
7560 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
7561 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
7562 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
7563 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
7564 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
7565 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
7566 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
7567 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
7568 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
7569 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
7570 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
7571 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
7573 — Berlin, 2014-02-20
7577 * logind has gained support for facilitating privileged input
7578 and drm device access for unprivileged clients. This work is
7579 useful to allow Wayland display servers (and similar
7580 programs, such as kmscon) to run under the user's ID and
7581 access input and drm devices which are normally
7582 protected. When this is used (and the kernel is new enough)
7583 logind will "mute" IO on the file descriptors passed to
7584 Wayland as long as it is in the background and "unmute" it
7585 if it returns into the foreground. This allows secure
7586 session switching without allowing background sessions to
7587 eavesdrop on input and display data. This also introduces
7588 session switching support if VT support is turned off in the
7589 kernel, and on seats that are not seat0.
7591 * A new kernel command line option luks.options= is understood
7592 now which allows specifying LUKS options for usage for LUKS
7593 encrypted partitions specified with luks.uuid=.
7595 * tmpfiles.d(5) snippets may now use specifier expansion in
7596 path names. More specifically %m, %b, %H, %v, are now
7597 replaced by the local machine id, boot id, hostname, and
7598 kernel version number.
7600 * A new tmpfiles.d(5) command "m" has been introduced which
7601 may be used to change the owner/group/access mode of a file
7602 or directory if it exists, but do nothing if it does not.
7604 * This release removes high-level support for the
7605 MemorySoftLimit= cgroup setting. The underlying kernel
7606 cgroup attribute memory.soft_limit= is currently badly
7607 designed and likely to be removed from the kernel API in its
7608 current form, hence we should not expose it for now.
7610 * The memory.use_hierarchy cgroup attribute is now enabled for
7611 all cgroups systemd creates in the memory cgroup
7612 hierarchy. This option is likely to be come the built-in
7613 default in the kernel anyway, and the non-hierarchical mode
7614 never made much sense in the intrinsically hierarchical
7617 * A new field _SYSTEMD_SLICE= is logged along with all journal
7618 messages containing the slice a message was generated
7619 from. This is useful to allow easy per-customer filtering of
7620 logs among other things.
7622 * systemd-journald will no longer adjust the group of journal
7623 files it creates to the "systemd-journal" group. Instead we
7624 rely on the journal directory to be owned by the
7625 "systemd-journal" group, and its setgid bit set, so that the
7626 kernel file system layer will automatically enforce that
7627 journal files inherit this group assignment. The reason for
7628 this change is that we cannot allow NSS look-ups from
7629 journald which would be necessary to resolve
7630 "systemd-journal" to a numeric GID, because this might
7631 create deadlocks if NSS involves synchronous queries to
7632 other daemons (such as nscd, or sssd) which in turn are
7633 logging clients of journald and might block on it, which
7634 would then dead lock. A tmpfiles.d(5) snippet included in
7635 systemd will make sure the setgid bit and group are
7636 properly set on the journal directory if it exists on every
7637 boot. However, we recommend adjusting it manually after
7638 upgrades too (or from RPM scriptlets), so that the change is
7639 not delayed until next reboot.
7641 * Backlight and random seed files in /var/lib/ have moved into
7642 the /var/lib/systemd/ directory, in order to centralize all
7643 systemd generated files in one directory.
7645 * Boot time performance measurements (as displayed by
7646 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
7647 performance information if that's available to determine how
7648 much time BIOS and boot loader initialization required. With
7649 a sufficiently new BIOS you hence no longer need to boot
7650 with Gummiboot to get access to such information.
7652 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
7653 Cristian Rodríguez, Dave Reisner, David Herrmann, David
7654 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
7655 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
7656 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
7657 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
7658 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7660 — Berlin, 2013-10-02
7664 * The Restart= option for services now understands a new
7665 on-watchdog setting, which will restart the service
7666 automatically if the service stops sending out watchdog keep
7667 alive messages (as configured with WatchdogSec=).
7669 * The getty generator (which is responsible for bringing up a
7670 getty on configured serial consoles) will no longer only
7671 start a getty on the primary kernel console but on all
7672 others, too. This makes the order in which console= is
7673 specified on the kernel command line less important.
7675 * libsystemd-logind gained a new sd_session_get_vt() call to
7676 retrieve the VT number of a session.
7678 * If the option "tries=0" is set for an entry of /etc/crypttab
7679 its passphrase is queried indefinitely instead of any
7680 maximum number of tries.
7682 * If a service with a configure PID file terminates its PID
7683 file will now be removed automatically if it still exists
7684 afterwards. This should put an end to stale PID files.
7686 * systemd-run will now also take relative binary path names
7687 for execution and no longer insists on absolute paths.
7689 * InaccessibleDirectories= and ReadOnlyDirectories= now take
7690 paths that are optionally prefixed with "-" to indicate that
7691 it should not be considered a failure if they do not exist.
7693 * journalctl -o (and similar commands) now understands a new
7694 output mode "short-precise", it is similar to "short" but
7695 shows timestamps with usec accuracy.
7697 * The option "discard" (as known from Debian) is now
7698 synonymous to "allow-discards" in /etc/crypttab. In fact,
7699 "discard" is preferred now (since it is easier to remember
7702 * Some licensing clean-ups were made, so that more code is now
7703 LGPL-2.1 licensed than before.
7705 * A minimal tool to save/restore the display backlight
7706 brightness across reboots has been added. It will store the
7707 backlight setting as late as possible at shutdown, and
7708 restore it as early as possible during reboot.
7710 * A logic to automatically discover and enable home and swap
7711 partitions on GPT disks has been added. With this in place
7712 /etc/fstab becomes optional for many setups as systemd can
7713 discover certain partitions located on the root disk
7714 automatically. Home partitions are recognized under their
7715 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
7716 partitions are recognized under their GPT type ID
7717 0657fd6da4ab43c484e50933c84b4f4f.
7719 * systemd will no longer pass any environment from the kernel
7720 or initrd to system services. If you want to set an
7721 environment for all services, do so via the kernel command
7722 line systemd.setenv= assignment.
7724 * The systemd-sysctl tool no longer natively reads the file
7725 /etc/sysctl.conf. If desired, the file should be symlinked
7726 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
7727 legacy support by a symlink rather than built-in code, it
7728 also makes the otherwise hidden order of application of the
7729 different files visible. (Note that this partly reverts to a
7730 pre-198 application order of sysctl knobs!)
7732 * The "systemctl set-log-level" and "systemctl dump" commands
7733 have been moved to systemd-analyze.
7735 * systemd-run learned the new --remain-after-exit switch,
7736 which causes the scope unit not to be cleaned up
7737 automatically after the process terminated.
7739 * tmpfiles learned a new --exclude-prefix= switch to exclude
7740 certain paths from operation.
7742 * journald will now automatically flush all messages to disk
7743 as soon as a message at the log level CRIT, ALERT or EMERG
7746 Contributions from: Andrew Cook, Brandon Philips, Christian
7747 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
7748 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
7749 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
7750 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
7751 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
7752 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
7753 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
7754 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
7755 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
7756 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
7757 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
7758 William Giokas, Zbigniew Jędrzejewski-Szmek
7760 — Berlin, 2013-09-13
7764 * The documentation has been updated to cover the various new
7765 concepts introduced with 205.
7767 * Unit files now understand the new %v specifier which
7768 resolves to the kernel version string as returned by "uname
7771 * systemctl now supports filtering the unit list output by
7772 load state, active state and sub state, using the new
7775 * "systemctl status" will now show the results of the
7776 condition checks (like ConditionPathExists= and similar) of
7777 the last start attempts of the unit. They are also logged to
7780 * "journalctl -b" may now be used to look for boot output of a
7781 specific boot. Try "journalctl -b -1" for the previous boot,
7782 but the syntax is substantially more powerful.
7784 * "journalctl --show-cursor" has been added which prints the
7785 cursor string the last shown log line. This may then be used
7786 with the new "journalctl --after-cursor=" switch to continue
7787 browsing logs from that point on.
7789 * "journalctl --force" may now be used to force regeneration
7792 * Creation of "dead" device nodes has been moved from udev
7793 into kmod and tmpfiles. Previously, udev would read the kmod
7794 databases to pre-generate dead device nodes based on meta
7795 information contained in kernel modules, so that these would
7796 be auto-loaded on access rather then at boot. As this
7797 does not really have much to do with the exposing actual
7798 kernel devices to userspace this has always been slightly
7799 alien in the udev codebase. Following the new scheme kmod
7800 will now generate a runtime snippet for tmpfiles from the
7801 module meta information and it now is tmpfiles' job to the
7802 create the nodes. This also allows overriding access and
7803 other parameters for the nodes using the usual tmpfiles
7804 facilities. As side effect this allows us to remove the
7805 CAP_SYS_MKNOD capability bit from udevd entirely.
7807 * logind's device ACLs may now be applied to these "dead"
7808 devices nodes too, thus finally allowing managed access to
7809 devices such as /dev/snd/sequencer without loading the
7810 backing module right-away.
7812 * A new RPM macro has been added that may be used to apply
7813 tmpfiles configuration during package installation.
7815 * systemd-detect-virt and ConditionVirtualization= now can
7816 detect User-Mode-Linux machines (UML).
7818 * journald will now implicitly log the effective capabilities
7819 set of processes in the message metadata.
7821 * systemd-cryptsetup has gained support for TrueCrypt volumes.
7823 * The initrd interface has been simplified (more specifically,
7824 support for passing performance data via environment
7825 variables and fsck results via files in /run has been
7826 removed). These features were non-essential, and are
7827 nowadays available in a much nicer way by having systemd in
7828 the initrd serialize its state and have the hosts systemd
7829 deserialize it again.
7831 * The udev "keymap" data files and tools to apply keyboard
7832 specific mappings of scan to key codes, and force-release
7833 scan code lists have been entirely replaced by a udev
7834 "keyboard" builtin and a hwdb data file.
7836 * systemd will now honour the kernel's "quiet" command line
7837 argument also during late shutdown, resulting in a
7838 completely silent shutdown when used.
7840 * There's now an option to control the SO_REUSEPORT socket
7841 option in .socket units.
7843 * Instance units will now automatically get a per-template
7844 subslice of system.slice unless something else is explicitly
7845 configured. For example, instances of sshd@.service will now
7846 implicitly be placed in system-sshd.slice rather than
7847 system.slice as before.
7849 * Test coverage support may now be enabled at build time.
7851 Contributions from: Dave Reisner, Frederic Crozat, Harald
7852 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
7853 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
7854 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
7855 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
7856 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
7857 Giokas, Zbigniew Jędrzejewski-Szmek
7859 — Berlin, 2013-07-23
7863 * Two new unit types have been introduced:
7865 Scope units are very similar to service units, however, are
7866 created out of pre-existing processes — instead of PID 1
7867 forking off the processes. By using scope units it is
7868 possible for system services and applications to group their
7869 own child processes (worker processes) in a powerful way
7870 which then maybe used to organize them, or kill them
7871 together, or apply resource limits on them.
7873 Slice units may be used to partition system resources in an
7874 hierarchical fashion and then assign other units to them. By
7875 default there are now three slices: system.slice (for all
7876 system services), user.slice (for all user sessions),
7877 machine.slice (for VMs and containers).
7879 Slices and scopes have been introduced primarily in
7880 context of the work to move cgroup handling to a
7881 single-writer scheme, where only PID 1
7882 creates/removes/manages cgroups.
7884 * There's a new concept of "transient" units. In contrast to
7885 normal units these units are created via an API at runtime,
7886 not from configuration from disk. More specifically this
7887 means it is now possible to run arbitrary programs as
7888 independent services, with all execution parameters passed
7889 in via bus APIs rather than read from disk. Transient units
7890 make systemd substantially more dynamic then it ever was,
7891 and useful as a general batch manager.
7893 * logind has been updated to make use of scope and slice units
7894 for managing user sessions. As a user logs in he will get
7895 his own private slice unit, to which all sessions are added
7896 as scope units. We also added support for automatically
7897 adding an instance of user@.service for the user into the
7898 slice. Effectively logind will no longer create cgroup
7899 hierarchies on its own now, it will defer entirely to PID 1
7900 for this by means of scope, service and slice units. Since
7901 user sessions this way become entities managed by PID 1
7902 the output of "systemctl" is now a lot more comprehensive.
7904 * A new mini-daemon "systemd-machined" has been added which
7905 may be used by virtualization managers to register local
7906 VMs/containers. nspawn has been updated accordingly, and
7907 libvirt will be updated shortly. machined will collect a bit
7908 of meta information about the VMs/containers, and assign
7909 them their own scope unit (see above). The collected
7910 meta-data is then made available via the "machinectl" tool,
7911 and exposed in "ps" and similar tools. machined/machinectl
7912 is compile-time optional.
7914 * As discussed earlier, the low-level cgroup configuration
7915 options ControlGroup=, ControlGroupModify=,
7916 ControlGroupPersistent=, ControlGroupAttribute= have been
7917 removed. Please use high-level attribute settings instead as
7918 well as slice units.
7920 * A new bus call SetUnitProperties() has been added to alter
7921 various runtime parameters of a unit. This is primarily
7922 useful to alter cgroup parameters dynamically in a nice way,
7923 but will be extended later on to make more properties
7924 modifiable at runtime. systemctl gained a new set-properties
7925 command that wraps this call.
7927 * A new tool "systemd-run" has been added which can be used to
7928 run arbitrary command lines as transient services or scopes,
7929 while configuring a number of settings via the command
7930 line. This tool is currently very basic, however already
7931 very useful. We plan to extend this tool to even allow
7932 queuing of execution jobs with time triggers from the
7933 command line, similar in fashion to "at".
7935 * nspawn will now inform the user explicitly that kernels with
7936 audit enabled break containers, and suggest the user to turn
7939 * Support for detecting the IMA and AppArmor security
7940 frameworks with ConditionSecurity= has been added.
7942 * journalctl gained a new "-k" switch for showing only kernel
7943 messages, mimicking dmesg output; in addition to "--user"
7944 and "--system" switches for showing only user's own logs
7947 * systemd-delta can now show information about drop-in
7948 snippets extending unit files.
7950 * libsystemd-bus has been substantially updated but is still
7951 not available as public API.
7953 * systemd will now look for the "debug" argument on the kernel
7954 command line and enable debug logging, similar to what
7955 "systemd.log_level=debug" already did before.
7957 * "systemctl set-default", "systemctl get-default" has been
7958 added to configure the default.target symlink, which
7959 controls what to boot into by default.
7961 * "systemctl set-log-level" has been added as a convenient
7962 way to raise and lower systemd logging threshold.
7964 * "systemd-analyze plot" will now show the time the various
7965 generators needed for execution, as well as information
7966 about the unit file loading.
7968 * libsystemd-journal gained a new sd_journal_open_files() call
7969 for opening specific journal files. journactl also gained a
7970 new switch to expose this new functionality. Previously we
7971 only supported opening all files from a directory, or all
7972 files from the system, as opening individual files only is
7973 racy due to journal file rotation.
7975 * systemd gained the new DefaultEnvironment= setting in
7976 /etc/systemd/system.conf to set environment variables for
7979 * If a privileged process logs a journal message with the
7980 OBJECT_PID= field set, then journald will automatically
7981 augment this with additional OBJECT_UID=, OBJECT_GID=,
7982 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
7983 system services want to log events about specific client
7984 processes. journactl/systemctl has been updated to make use
7985 of this information if all log messages regarding a specific
7988 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
7989 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
7990 Reisner, David Coppa, David King, David Strauss, Eelco
7991 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
7992 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
7993 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
7994 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
7995 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
7996 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
7997 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
7998 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
7999 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
8000 Łukasz Stelmach, 장동준
8004 * The Python bindings gained some minimal support for the APIs
8005 exposed by libsystemd-logind.
8007 * ConditionSecurity= gained support for detecting SMACK. Since
8008 this condition already supports SELinux and AppArmor we only
8009 miss IMA for this. Patches welcome!
8011 Contributions from: Karol Lewandowski, Lennart Poettering,
8012 Zbigniew Jędrzejewski-Szmek
8016 * systemd-nspawn will now create /etc/resolv.conf if
8017 necessary, before bind-mounting the host's file onto it.
8019 * systemd-nspawn will now store meta information about a
8020 container on the container's cgroup as extended attribute
8021 fields, including the root directory.
8023 * The cgroup hierarchy has been reworked in many ways. All
8024 objects any of the components systemd creates in the cgroup
8025 tree are now suffixed. More specifically, user sessions are
8026 now placed in cgroups suffixed with ".session", users in
8027 cgroups suffixed with ".user", and nspawn containers in
8028 cgroups suffixed with ".nspawn". Furthermore, all cgroup
8029 names are now escaped in a simple scheme to avoid collision
8030 of userspace object names with kernel filenames. This work
8031 is preparation for making these objects relocatable in the
8032 cgroup tree, in order to allow easy resource partitioning of
8033 these objects without causing naming conflicts.
8035 * systemctl list-dependencies gained the new switches
8036 --plain, --reverse, --after and --before.
8038 * systemd-inhibit now shows the process name of processes that
8039 have taken an inhibitor lock.
8041 * nss-myhostname will now also resolve "localhost"
8042 implicitly. This makes /etc/hosts an optional file and
8043 nicely handles that on IPv6 ::1 maps to both "localhost" and
8046 * libsystemd-logind.so gained a new call
8047 sd_get_machine_names() to enumerate running containers and
8048 VMs (currently only supported by very new libvirt and
8049 nspawn). sd_login_monitor can now be used to watch
8050 VMs/containers coming and going.
8052 * .include is not allowed recursively anymore, and only in
8053 unit files. Usually it is better to use drop-in snippets in
8054 .d/*.conf anyway, as introduced with systemd 198.
8056 * systemd-analyze gained a new "critical-chain" command that
8057 determines the slowest chain of units run during system
8058 boot-up. It is very useful for tracking down where
8059 optimizing boot time is the most beneficial.
8061 * systemd will no longer allow manipulating service paths in
8062 the name=systemd:/system cgroup tree using ControlGroup= in
8063 units. (But is still fine with it in all other dirs.)
8065 * There's a new systemd-nspawn@.service service file that may
8066 be used to easily run nspawn containers as system
8067 services. With the container's root directory in
8068 /var/lib/container/foobar it is now sufficient to run
8069 "systemctl start systemd-nspawn@foobar.service" to boot it.
8071 * systemd-cgls gained a new parameter "--machine" to list only
8072 the processes within a certain container.
8074 * ConditionSecurity= now can check for "apparmor". We still
8075 are lacking checks for SMACK and IMA for this condition
8076 check though. Patches welcome!
8078 * A new configuration file /etc/systemd/sleep.conf has been
8079 added that may be used to configure which kernel operation
8080 systemd is supposed to execute when "suspend", "hibernate"
8081 or "hybrid-sleep" is requested. This makes the new kernel
8082 "freeze" state accessible to the user.
8084 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
8085 the passed argument if applicable.
8087 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
8088 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
8089 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
8090 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
8091 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
8092 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
8093 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
8098 * The output of 'systemctl list-jobs' got some polishing. The
8099 '--type=' argument may now be passed more than once. A new
8100 command 'systemctl list-sockets' has been added which shows
8101 a list of kernel sockets systemd is listening on with the
8102 socket units they belong to, plus the units these socket
8105 * The experimental libsystemd-bus library got substantial
8106 updates to work in conjunction with the (also experimental)
8107 kdbus kernel project. It works well enough to exchange
8108 messages with some sophistication. Note that kdbus is not
8109 ready yet, and the library is mostly an elaborate test case
8110 for now, and not installable.
8112 * systemd gained a new unit 'systemd-static-nodes.service'
8113 that generates static device nodes earlier during boot, and
8114 can run in conjunction with udev.
8116 * libsystemd-login gained a new call sd_pid_get_user_unit()
8117 to retrieve the user systemd unit a process is running
8118 in. This is useful for systems where systemd is used as
8121 * systemd-nspawn now places all containers in the new /machine
8122 top-level cgroup directory in the name=systemd
8123 hierarchy. libvirt will soon do the same, so that we get a
8124 uniform separation of /system, /user and /machine for system
8125 services, user processes and containers/virtual
8126 machines. This new cgroup hierarchy is also useful to stick
8127 stable names to specific container instances, which can be
8128 recognized later this way (this name may be controlled
8129 via systemd-nspawn's new -M switch). libsystemd-login also
8130 gained a new call sd_pid_get_machine_name() to retrieve the
8131 name of the container/VM a specific process belongs to.
8133 * bootchart can now store its data in the journal.
8135 * libsystemd-journal gained a new call
8136 sd_journal_add_conjunction() for AND expressions to the
8137 matching logic. This can be used to express more complex
8138 logical expressions.
8140 * journactl can now take multiple --unit= and --user-unit=
8143 * The cryptsetup logic now understands the "luks.key=" kernel
8144 command line switch for specifying a file to read the
8145 decryption key from. Also, if a configured key file is not
8146 found the tool will now automatically fall back to prompting
8149 * Python systemd.journal module was updated to wrap recently
8150 added functions from libsystemd-journal. The interface was
8151 changed to bring the low level interface in s.j._Reader
8152 closer to the C API, and the high level interface in
8153 s.j.Reader was updated to wrap and convert all data about
8156 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
8157 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
8158 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
8159 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
8160 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
8161 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8165 * journalctl --update-catalog now understands a new --root=
8166 option to operate on catalogs found in a different root
8169 * During shutdown after systemd has terminated all running
8170 services a final killing loop kills all remaining left-over
8171 processes. We will now print the name of these processes
8172 when we send SIGKILL to them, since this usually indicates a
8175 * If /etc/crypttab refers to password files stored on
8176 configured mount points automatic dependencies will now be
8177 generated to ensure the specific mount is established first
8178 before the key file is attempted to be read.
8180 * 'systemctl status' will now show information about the
8181 network sockets a socket unit is listening on.
8183 * 'systemctl status' will also shown information about any
8184 drop-in configuration file for units. (Drop-In configuration
8185 files in this context are files such as
8186 /etc/systemd/systemd/foobar.service.d/*.conf)
8188 * systemd-cgtop now optionally shows summed up CPU times of
8189 cgroups. Press '%' while running cgtop to switch between
8190 percentage and absolute mode. This is useful to determine
8191 which cgroups use up the most CPU time over the entire
8192 runtime of the system. systemd-cgtop has also been updated
8193 to be 'pipeable' for processing with further shell tools.
8195 * 'hostnamectl set-hostname' will now allow setting of FQDN
8198 * The formatting and parsing of time span values has been
8199 changed. The parser now understands fractional expressions
8200 such as "5.5h". The formatter will now output fractional
8201 expressions for all time spans under 1min, i.e. "5.123456s"
8202 rather than "5s 123ms 456us". For time spans under 1s
8203 millisecond values are shown, for those under 1ms
8204 microsecond values are shown. This should greatly improve
8205 all time-related output of systemd.
8207 * libsystemd-login and libsystemd-journal gained new
8208 functions for querying the poll() events mask and poll()
8209 timeout value for integration into arbitrary event
8212 * localectl gained the ability to list available X11 keymaps
8213 (models, layouts, variants, options).
8215 * 'systemd-analyze dot' gained the ability to filter for
8216 specific units via shell-style globs, to create smaller,
8217 more useful graphs. I.e. it is now possible to create simple
8218 graphs of all the dependencies between only target units, or
8219 of all units that Avahi has dependencies with.
8221 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
8222 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
8223 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
8224 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
8225 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
8226 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
8227 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
8231 * The boot-time readahead implementation for rotating media
8232 will now read the read-ahead data in multiple passes which
8233 consist of all read requests made in equidistant time
8234 intervals. This means instead of strictly reading read-ahead
8235 data in its physical order on disk we now try to find a
8236 middle ground between physical and access time order.
8238 * /etc/os-release files gained a new BUILD_ID= field for usage
8239 on operating systems that provide continuous builds of OS
8242 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
8243 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
8244 William Douglas, Zbigniew Jędrzejewski-Szmek
8248 * systemd-python gained an API exposing libsystemd-daemon.
8250 * The SMACK setup logic gained support for uploading CIPSO
8253 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
8254 ReadOnlyDirectories= and InaccessibleDirectories= has
8255 changed. The private /tmp and /var/tmp directories are now
8256 shared by all processes of a service (which means
8257 ExecStartPre= may now leave data in /tmp that ExecStart= of
8258 the same service can still access). When a service is
8259 stopped its temporary directories are immediately deleted
8260 (normal clean-up with tmpfiles is still done in addition to
8263 * By default, systemd will now set a couple of sysctl
8264 variables in the kernel: the safe sysrq options are turned
8265 on, IP route verification is turned on, and source routing
8266 disabled. The recently added hardlink and softlink
8267 protection of the kernel is turned on. These settings should
8268 be reasonably safe, and good defaults for all new systems.
8270 * The predictable network naming logic may now be turned off
8271 with a new kernel command line switch: net.ifnames=0.
8273 * A new libsystemd-bus module has been added that implements a
8274 pretty complete D-Bus client library. For details see:
8276 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
8278 * journald will now explicitly flush the journal files to disk
8279 at the latest 5min after each write. The file will then also
8280 be marked offline until the next write. This should increase
8281 reliability in case of a crash. The synchronization delay
8282 can be configured via SyncIntervalSec= in journald.conf.
8284 * There's a new remote-fs-setup.target unit that can be used
8285 to pull in specific services when at least one remote file
8286 system is to be mounted.
8288 * There are new targets timers.target and paths.target as
8289 canonical targets to pull user timer and path units in
8290 from. This complements sockets.target with a similar
8291 purpose for socket units.
8293 * libudev gained a new call udev_device_set_attribute_value()
8294 to set sysfs attributes of a device.
8296 * The udev daemon now sets the default number of worker
8297 processes executed in parallel based on the number of available
8298 CPUs instead of the amount of available RAM. This is supposed
8299 to provide a more reliable default and limit a too aggressive
8300 parallelism for setups with 1000s of devices connected.
8302 Contributions from: Auke Kok, Colin Walters, Cristian
8303 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
8304 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
8305 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
8306 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
8307 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
8308 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
8309 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
8310 Zbigniew Jędrzejewski-Szmek
8314 * Configuration of unit files may now be extended via drop-in
8315 files without having to edit/override the unit files
8316 themselves. More specifically, if the administrator wants to
8317 change one value for a service file foobar.service he can
8318 now do so by dropping in a configuration snippet into
8319 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
8320 will load all these snippets and apply them on top of the
8321 main unit configuration file, possibly extending or
8322 overriding its settings. Using these drop-in snippets is
8323 generally nicer than the two earlier options for changing
8324 unit files locally: copying the files from
8325 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
8326 them there; or creating a new file in /etc/systemd/system/
8327 that incorporates the original one via ".include". Drop-in
8328 snippets into these .d/ directories can be placed in any
8329 directory systemd looks for units in, and the usual
8330 overriding semantics between /usr/lib, /etc and /run apply
8333 * Most unit file settings which take lists of items can now be
8334 reset by assigning the empty string to them. For example,
8335 normally, settings such as Environment=FOO=BAR append a new
8336 environment variable assignment to the environment block,
8337 each time they are used. By assigning Environment= the empty
8338 string the environment block can be reset to empty. This is
8339 particularly useful with the .d/*.conf drop-in snippets
8340 mentioned above, since this adds the ability to reset list
8341 settings from vendor unit files via these drop-ins.
8343 * systemctl gained a new "list-dependencies" command for
8344 listing the dependencies of a unit recursively.
8346 * Inhibitors are now honored and listed by "systemctl
8347 suspend", "systemctl poweroff" (and similar) too, not only
8348 GNOME. These commands will also list active sessions by
8351 * Resource limits (as exposed by the various control group
8352 controllers) can now be controlled dynamically at runtime
8353 for all units. More specifically, you can now use a command
8354 like "systemctl set-cgroup-attr foobar.service cpu.shares
8355 2000" to alter the CPU shares a specific service gets. These
8356 settings are stored persistently on disk, and thus allow the
8357 administrator to easily adjust the resource usage of
8358 services with a few simple commands. This dynamic resource
8359 management logic is also available to other programs via the
8360 bus. Almost any kernel cgroup attribute and controller is
8363 * systemd-vconsole-setup will now copy all font settings to
8364 all allocated VTs, where it previously applied them only to
8367 * libsystemd-login gained the new sd_session_get_tty() API
8370 * This release drops support for a few legacy or
8371 distribution-specific LSB facility names when parsing init
8372 scripts: $x-display-manager, $mail-transfer-agent,
8373 $mail-transport-agent, $mail-transfer-agent, $smtp,
8374 $null. Also, the mail-transfer-agent.target unit backing
8375 this has been removed. Distributions which want to retain
8376 compatibility with this should carry the burden for
8377 supporting this themselves and patch support for these back
8378 in, if they really need to. Also, the facilities $syslog and
8379 $local_fs are now ignored, since systemd does not support
8380 early-boot LSB init scripts anymore, and these facilities
8381 are implied anyway for normal services. syslog.target has
8384 * There are new bus calls on PID1's Manager object for
8385 cancelling jobs, and removing snapshot units. Previously,
8386 both calls were only available on the Job and Snapshot
8389 * systemd-journal-gatewayd gained SSL support.
8391 * The various "environment" files, such as /etc/locale.conf
8392 now support continuation lines with a backslash ("\") as
8393 last character in the line, similarly in style (but different)
8394 to how this is supported in shells.
8396 * For normal user processes the _SYSTEMD_USER_UNIT= field is
8397 now implicitly appended to every log entry logged. systemctl
8398 has been updated to filter by this field when operating on a
8399 user systemd instance.
8401 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
8402 CAP_AUDIT_CONTROL capabilities to the capabilities set for
8403 the container. This makes it easier to boot unmodified
8404 Fedora systems in a container, which however still requires
8405 audit=0 to be passed on the kernel command line. Auditing in
8406 kernel and userspace is unfortunately still too broken in
8407 context of containers, hence we recommend compiling it out
8408 of the kernel or using audit=0. Hopefully this will be fixed
8409 one day for good in the kernel.
8411 * nspawn gained the new --bind= and --bind-ro= parameters to
8412 bind mount specific directories from the host into the
8415 * nspawn will now mount its own devpts file system instance
8416 into the container, in order not to leak pty devices from
8417 the host into the container.
8419 * systemd will now read the firmware boot time performance
8420 information from the EFI variables, if the used boot loader
8421 supports this, and takes it into account for boot performance
8422 analysis via "systemd-analyze". This is currently supported
8423 only in conjunction with Gummiboot, but could be supported
8424 by other boot loaders too. For details see:
8426 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
8428 * A new generator has been added that automatically mounts the
8429 EFI System Partition (ESP) to /boot, if that directory
8430 exists, is empty, and no other file system has been
8431 configured to be mounted there.
8433 * logind will now send out PrepareForSleep(false) out
8434 unconditionally, after coming back from suspend. This may be
8435 used by applications as asynchronous notification for
8436 system resume events.
8438 * "systemctl unlock-sessions" has been added, that allows
8439 unlocking the screens of all user sessions at once, similar
8440 to how "systemctl lock-sessions" already locked all users
8441 sessions. This is backed by a new D-Bus call UnlockSessions().
8443 * "loginctl seat-status" will now show the master device of a
8444 seat. (i.e. the device of a seat that needs to be around for
8445 the seat to be considered available, usually the graphics
8448 * tmpfiles gained a new "X" line type, that allows
8449 configuration of files and directories (with wildcards) that
8450 shall be excluded from automatic cleanup ("aging").
8452 * udev default rules set the device node permissions now only
8453 at "add" events, and do not change them any longer with a
8454 later "change" event.
8456 * The log messages for lid events and power/sleep keypresses
8457 now carry a message ID.
8459 * We now have a substantially larger unit test suite, but this
8460 continues to be work in progress.
8462 * udevadm hwdb gained a new --root= parameter to change the
8463 root directory to operate relative to.
8465 * logind will now issue a background sync() request to the kernel
8466 early at shutdown, so that dirty buffers are flushed to disk early
8467 instead of at the last moment, in order to optimize shutdown
8470 * A new bootctl tool has been added that is an interface for
8471 certain boot loader operations. This is currently a preview
8472 and is likely to be extended into a small mechanism daemon
8473 like timedated, localed, hostnamed, and can be used by
8474 graphical UIs to enumerate available boot options, and
8475 request boot into firmware operations.
8477 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
8478 the rest of the package. It also has been updated to work
8479 correctly in initrds.
8481 * polkit previously has been runtime optional, and is now also
8482 compile time optional via a configure switch.
8484 * systemd-analyze has been reimplemented in C. Also "systemctl
8485 dot" has moved into systemd-analyze.
8487 * "systemctl status" with no further parameters will now print
8488 the status of all active or failed units.
8490 * Operations such as "systemctl start" can now be executed
8491 with a new mode "--irreversible" which may be used to queue
8492 operations that cannot accidentally be reversed by a later
8493 job queuing. This is by default used to make shutdown
8494 requests more robust.
8496 * The Python API of systemd now gained a new module for
8497 reading journal files.
8499 * A new tool kernel-install has been added that can install
8500 kernel images according to the Boot Loader Specification:
8502 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
8504 * Boot time console output has been improved to provide
8505 animated boot time output for hanging jobs.
8507 * A new tool systemd-activate has been added which can be used
8508 to test socket activation with, directly from the command
8509 line. This should make it much easier to test and debug
8510 socket activation in daemons.
8512 * journalctl gained a new "--reverse" (or -r) option to show
8513 journal output in reverse order (i.e. newest line first).
8515 * journalctl gained a new "--pager-end" (or -e) option to jump
8516 to immediately jump to the end of the journal in the
8517 pager. This is only supported in conjunction with "less".
8519 * journalctl gained a new "--user-unit=" option, that works
8520 similarly to "--unit=" but filters for user units rather than
8523 * A number of unit files to ease adoption of systemd in
8524 initrds has been added. This moves some minimal logic from
8525 the various initrd implementations into systemd proper.
8527 * The journal files are now owned by a new group
8528 "systemd-journal", which exists specifically to allow access
8529 to the journal, and nothing else. Previously, we used the
8530 "adm" group for that, which however possibly covers more
8531 than just journal/log file access. This new group is now
8532 already used by systemd-journal-gatewayd to ensure this
8533 daemon gets access to the journal files and as little else
8534 as possible. Note that "make install" will also set FS ACLs
8535 up for /var/log/journal to give "adm" and "wheel" read
8536 access to it, in addition to "systemd-journal" which owns
8537 the journal files. We recommend that packaging scripts also
8538 add read access to "adm" + "wheel" to /var/log/journal, and
8539 all existing/future journal files. To normal users and
8540 administrators little changes, however packagers need to
8541 ensure to create the "systemd-journal" system group at
8542 package installation time.
8544 * The systemd-journal-gatewayd now runs as unprivileged user
8545 systemd-journal-gateway:systemd-journal-gateway. Packaging
8546 scripts need to create these system user/group at
8549 * timedated now exposes a new boolean property CanNTP that
8550 indicates whether a local NTP service is available or not.
8552 * systemd-detect-virt will now also detect xen PVs
8554 * The pstore file system is now mounted by default, if it is
8557 * In addition to the SELinux and IMA policies we will now also
8558 load SMACK policies at early boot.
8560 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
8561 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
8562 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
8563 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
8564 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
8565 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
8566 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
8567 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
8568 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
8569 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
8570 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
8571 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
8572 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
8573 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
8577 * Timer units now support calendar time events in addition to
8578 monotonic time events. That means you can now trigger a unit
8579 based on a calendar time specification such as "Thu,Fri
8580 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
8581 or fifth day of any month of the year 2013, given that it is
8582 a thursday or friday. This brings timer event support
8583 considerably closer to cron's capabilities. For details on
8584 the supported calendar time specification language see
8587 * udev now supports a number of different naming policies for
8588 network interfaces for predictable names, and a combination
8589 of these policies is now the default. Please see this wiki
8590 document for details:
8592 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8594 * Auke Kok's bootchart implementation has been added to the
8595 systemd tree. It is an optional component that can graph the
8596 boot in quite some detail. It is one of the best bootchart
8597 implementations around and minimal in its code and
8600 * nss-myhostname has been integrated into the systemd source
8601 tree. nss-myhostname guarantees that the local hostname
8602 always stays resolvable via NSS. It has been a weak
8603 requirement of systemd-hostnamed since a long time, and
8604 since its code is actually trivial we decided to just
8605 include it in systemd's source tree. It can be turned off
8606 with a configure switch.
8608 * The read-ahead logic is now capable of properly detecting
8609 whether a btrfs file system is on SSD or rotating media, in
8610 order to optimize the read-ahead scheme. Previously, it was
8611 only capable of detecting this on traditional file systems
8614 * In udev, additional device properties are now read from the
8615 IAB in addition to the OUI database. Also, Bluetooth company
8616 identities are attached to the devices as well.
8618 * In service files %U may be used as specifier that is
8619 replaced by the configured user name of the service.
8621 * nspawn may now be invoked without a controlling TTY. This
8622 makes it suitable for invocation as its own service. This
8623 may be used to set up a simple containerized server system
8624 using only core OS tools.
8626 * systemd and nspawn can now accept socket file descriptors
8627 when they are started for socket activation. This enables
8628 implementation of socket activated nspawn
8629 containers. i.e. think about autospawning an entire OS image
8630 when the first SSH or HTTP connection is received. We expect
8631 that similar functionality will also be added to libvirt-lxc
8634 * journalctl will now suppress ANSI color codes when
8635 presenting log data.
8637 * systemctl will no longer show control group information for
8638 a unit if the control group is empty anyway.
8640 * logind can now automatically suspend/hibernate/shutdown the
8643 * /etc/machine-info and hostnamed now also expose the chassis
8644 type of the system. This can be used to determine whether
8645 the local system is a laptop, desktop, handset or
8646 tablet. This information may either be configured by the
8647 user/vendor or is automatically determined from ACPI and DMI
8648 information if possible.
8650 * A number of polkit actions are now bound together with "imply"
8651 rules. This should simplify creating UIs because many actions
8652 will now authenticate similar ones as well.
8654 * Unit files learnt a new condition ConditionACPower= which
8655 may be used to conditionalize a unit depending on whether an
8656 AC power source is connected or not, of whether the system
8657 is running on battery power.
8659 * systemctl gained a new "is-failed" verb that may be used in
8660 shell scripts and suchlike to check whether a specific unit
8661 is in the "failed" state.
8663 * The EnvironmentFile= setting in unit files now supports file
8664 globbing, and can hence be used to easily read a number of
8665 environment files at once.
8667 * systemd will no longer detect and recognize specific
8668 distributions. All distribution-specific #ifdeffery has been
8669 removed, systemd is now fully generic and
8670 distribution-agnostic. Effectively, not too much is lost as
8671 a lot of the code is still accessible via explicit configure
8672 switches. However, support for some distribution specific
8673 legacy configuration file formats has been dropped. We
8674 recommend distributions to simply adopt the configuration
8675 files everybody else uses now and convert the old
8676 configuration from packaging scripts. Most distributions
8677 already did that. If that's not possible or desirable,
8678 distributions are welcome to forward port the specific
8679 pieces of code locally from the git history.
8681 * When logging a message about a unit systemd will now always
8682 log the unit name in the message meta data.
8684 * localectl will now also discover system locale data that is
8685 not stored in locale archives, but directly unpacked.
8687 * logind will no longer unconditionally use framebuffer
8688 devices as seat masters, i.e. as devices that are required
8689 to be existing before a seat is considered preset. Instead,
8690 it will now look for all devices that are tagged as
8691 "seat-master" in udev. By default, framebuffer devices will
8692 be marked as such, but depending on local systems, other
8693 devices might be marked as well. This may be used to
8694 integrate graphics cards using closed source drivers (such
8695 as NVidia ones) more nicely into logind. Note however, that
8696 we recommend using the open source NVidia drivers instead,
8697 and no udev rules for the closed-source drivers will be
8698 shipped from us upstream.
8700 Contributions from: Adam Williamson, Alessandro Crismani, Auke
8701 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
8702 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
8703 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
8704 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
8705 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
8706 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
8707 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
8708 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
8709 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
8710 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
8711 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
8716 * udev gained support for loading additional device properties
8717 from an indexed database that is keyed by vendor/product IDs
8718 and similar device identifiers. For the beginning this
8719 "hwdb" is populated with data from the well-known PCI and
8720 USB database, but also includes PNP, ACPI and OID data. In
8721 the longer run this indexed database shall grow into
8722 becoming the one central database for non-essential
8723 userspace device metadata. Previously, data from the PCI/USB
8724 database was only attached to select devices, since the
8725 lookup was a relatively expensive operation due to O(n) time
8726 complexity (with n being the number of entries in the
8727 database). Since this is now O(1), we decided to add in this
8728 data for all devices where this is available, by
8729 default. Note that the indexed database needs to be rebuilt
8730 when new data files are installed. To achieve this you need
8731 to update your packaging scripts to invoke "udevadm hwdb
8732 --update" after installation of hwdb data files. For
8733 RPM-based distributions we introduced the new
8734 %udev_hwdb_update macro for this purpose.
8736 * The Journal gained support for the "Message Catalog", an
8737 indexed database to link up additional information with
8738 journal entries. For further details please check:
8740 https://www.freedesktop.org/wiki/Software/systemd/catalog
8742 The indexed message catalog database also needs to be
8743 rebuilt after installation of message catalog files. Use
8744 "journalctl --update-catalog" for this. For RPM-based
8745 distributions we introduced the %journal_catalog_update
8746 macro for this purpose.
8748 * The Python Journal bindings gained support for the standard
8749 Python logging framework.
8751 * The Journal API gained new functions for checking whether
8752 the underlying file system of a journal file is capable of
8753 properly reporting file change notifications, or whether
8754 applications that want to reflect journal changes "live"
8755 need to recheck journal files continuously in appropriate
8758 * It is now possible to set the "age" field for tmpfiles
8759 entries to 0, indicating that files matching this entry
8760 shall always be removed when the directories are cleaned up.
8762 * coredumpctl gained a new "gdb" verb which invokes gdb
8763 right-away on the selected coredump.
8765 * There's now support for "hybrid sleep" on kernels that
8766 support this, in addition to "suspend" and "hibernate". Use
8767 "systemctl hybrid-sleep" to make use of this.
8769 * logind's HandleSuspendKey= setting (and related settings)
8770 now gained support for a new "lock" setting to simply
8771 request the screen lock on all local sessions, instead of
8772 actually executing a suspend or hibernation.
8774 * systemd will now mount the EFI variables file system by
8777 * Socket units now gained support for configuration of the
8778 SMACK security label.
8780 * timedatectl will now output the time of the last and next
8781 daylight saving change.
8783 * We dropped support for various legacy and distro-specific
8784 concepts, such as insserv, early-boot SysV services
8785 (i.e. those for non-standard runlevels such as 'b' or 'S')
8786 or ArchLinux /etc/rc.conf support. We recommend the
8787 distributions who still need support this to either continue
8788 to maintain the necessary patches downstream, or find a
8789 different solution. (Talk to us if you have questions!)
8791 * Various systemd components will now bypass polkit checks for
8792 root and otherwise handle properly if polkit is not found to
8793 be around. This should fix most issues for polkit-less
8794 systems. Quite frankly this should have been this way since
8795 day one. It is absolutely our intention to make systemd work
8796 fine on polkit-less systems, and we consider it a bug if
8797 something does not work as it should if polkit is not around.
8799 * For embedded systems it is now possible to build udev and
8800 systemd without blkid and/or kmod support.
8802 * "systemctl switch-root" is now capable of switching root
8803 more than once. I.e. in addition to transitions from the
8804 initrd to the host OS it is now possible to transition to
8805 further OS images from the host. This is useful to implement
8806 offline updating tools.
8808 * Various other additions have been made to the RPM macros
8809 shipped with systemd. Use %udev_rules_update() after
8810 installing new udev rules files. %_udevhwdbdir,
8811 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
8812 %_sysctldir are now available which resolve to the right
8813 directories for packages to place various data files in.
8815 * journalctl gained the new --full switch (in addition to
8816 --all, to disable ellipsation for long messages.
8818 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
8819 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
8820 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
8821 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
8822 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
8823 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
8824 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
8825 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
8826 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
8830 * journalctl gained new --since= and --until= switches to
8831 filter by time. It also now supports nice filtering for
8832 units via --unit=/-u.
8834 * Type=oneshot services may use ExecReload= and do the
8837 * The journal daemon now supports time-based rotation and
8838 vacuuming, in addition to the usual disk-space based
8841 * The journal will now index the available field values for
8842 each field name. This enables clients to show pretty drop
8843 downs of available match values when filtering. The bash
8844 completion of journalctl has been updated
8845 accordingly. journalctl gained a new switch -F to list all
8846 values a certain field takes in the journal database.
8848 * More service events are now written as structured messages
8849 to the journal, and made recognizable via message IDs.
8851 * The timedated, localed and hostnamed mini-services which
8852 previously only provided support for changing time, locale
8853 and hostname settings from graphical DEs such as GNOME now
8854 also have a minimal (but very useful) text-based client
8855 utility each. This is probably the nicest way to changing
8856 these settings from the command line now, especially since
8857 it lists available options and is fully integrated with bash
8860 * There's now a new tool "systemd-coredumpctl" to list and
8861 extract coredumps from the journal.
8863 * We now install a README each in /var/log/ and
8864 /etc/rc.d/init.d explaining where the system logs and init
8865 scripts went. This hopefully should help folks who go to
8866 that dirs and look into the otherwise now empty void and
8867 scratch their heads.
8869 * When user-services are invoked (by systemd --user) the
8870 $MANAGERPID env var is set to the PID of systemd.
8872 * SIGRTMIN+24 when sent to a --user instance will now result
8873 in immediate termination of systemd.
8875 * gatewayd received numerous feature additions such as a
8876 "follow" mode, for live syncing and filtering.
8878 * browse.html now allows filtering and showing detailed
8879 information on specific entries. Keyboard navigation and
8880 mouse screen support has been added.
8882 * gatewayd/journalctl now supports HTML5/JSON
8883 Server-Sent-Events as output.
8885 * The SysV init script compatibility logic will now
8886 heuristically determine whether a script supports the
8887 "reload" verb, and only then make this available as
8890 * "systemctl status --follow" has been removed, use "journalctl
8893 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
8894 have been removed since they are hardly useful to be
8897 * And I'd like to take the opportunity to specifically mention
8898 Zbigniew for his great contributions. Zbigniew, you rock!
8900 Contributions from: Andrew Eikum, Christian Hesse, Colin
8901 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
8902 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
8903 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
8904 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
8905 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
8906 Jędrzejewski-Szmek, Сковорода Никита Андреевич
8910 * If /etc/vconsole.conf is non-existent or empty we will no
8911 longer load any console font or key map at boot by
8912 default. Instead the kernel defaults will be left
8913 intact. This is definitely the right thing to do, as no
8914 configuration should mean no configuration, and hard-coding
8915 font names that are different on all archs is probably a bad
8916 idea. Also, the kernel default key map and font should be
8917 good enough for most cases anyway, and mostly identical to
8918 the userspace fonts/key maps we previously overloaded them
8919 with. If distributions want to continue to default to a
8920 non-kernel font or key map they should ship a default
8921 /etc/vconsole.conf with the appropriate contents.
8923 Contributions from: Colin Walters, Daniel J Walsh, Dave
8924 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
8925 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8929 * journalctl gained a new --cursor= switch to show entries
8930 starting from the specified location in the journal.
8932 * We now enforce a size limit on journal entry fields exported
8933 with "-o json" in journalctl. Fields larger than 4K will be
8934 assigned null. This can be turned off with --all.
8936 * An (optional) journal gateway daemon is now available as
8937 "systemd-journal-gatewayd.service". This service provides
8938 access to the journal via HTTP and JSON. This functionality
8939 will be used to implement live log synchronization in both
8940 pull and push modes, but has various other users too, such
8941 as easy log access for debugging of embedded devices. Right
8942 now it is already useful to retrieve the journal via HTTP:
8944 # systemctl start systemd-journal-gatewayd.service
8945 # wget http://localhost:19531/entries
8947 This will download the journal contents in a
8948 /var/log/messages compatible format. The same as JSON:
8950 # curl -H"Accept: application/json" http://localhost:19531/entries
8952 This service is also accessible via a web browser where a
8953 single static HTML5 app is served that uses the JSON logic
8954 to enable the user to do some basic browsing of the
8955 journal. This will be extended later on. Here's an example
8956 screenshot of this app in its current state:
8958 http://0pointer.de/public/journal-gatewayd
8960 Contributions from: Kay Sievers, Lennart Poettering, Robert
8961 Milasan, Tom Gundersen
8965 * The bash completion logic is now available for journalctl
8968 * We do not mount the "cpuset" controller anymore together with
8969 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
8970 started if no parameters are assigned to it. "cpuset" hence
8971 broke code that assumed it could create "cpu" groups and
8974 * journalctl -f will now subscribe to terminal size changes,
8975 and line break accordingly.
8977 Contributions from: Dave Reisner, Kay Sievers, Lennart
8978 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
8982 * nspawn will now create a symlink /etc/localtime in the
8983 container environment, copying the host's timezone
8984 setting. Previously this has been done via a bind mount, but
8985 since symlinks cannot be bind mounted this has now been
8986 changed to create/update the appropriate symlink.
8988 * journalctl -n's line number argument is now optional, and
8989 will default to 10 if omitted.
8991 * journald will now log the maximum size the journal files may
8992 take up on disk. This is particularly useful if the default
8993 built-in logic of determining this parameter from the file
8994 system size is used. Use "systemctl status
8995 systemd-journald.service" to see this information.
8997 * The multi-seat X wrapper tool has been stripped down. As X
8998 is now capable of enumerating graphics devices via udev in a
8999 seat-aware way the wrapper is not strictly necessary
9000 anymore. A stripped down temporary stop-gap is still shipped
9001 until the upstream display managers have been updated to
9002 fully support the new X logic. Expect this wrapper to be
9003 removed entirely in one of the next releases.
9005 * HandleSleepKey= in logind.conf has been split up into
9006 HandleSuspendKey= and HandleHibernateKey=. The old setting
9007 is not available anymore. X11 and the kernel are
9008 distinguishing between these keys and we should too. This
9009 also means the inhibition lock for these keys has been split
9012 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
9013 Poettering, Lukas Nykryn, Václav Pavlín
9017 * Whenever a unit changes state we will now log this to the
9018 journal and show along the unit's own log output in
9021 * ConditionPathIsMountPoint= can now properly detect bind
9022 mount points too. (Previously, a bind mount of one file
9023 system to another place in the same file system could not be
9024 detected as mount, since they shared struct stat's st_dev
9027 * We will now mount the cgroup controllers cpu, cpuacct,
9028 cpuset and the controllers net_cls, net_prio together by
9031 * nspawn containers will now have a virtualized boot
9032 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
9033 over with a randomized ID at container initialization). This
9034 has the effect of making "journalctl -b" do the right thing
9037 * The JSON output journal serialization has been updated not
9038 to generate "endless" list objects anymore, but rather one
9039 JSON object per line. This is more in line how most JSON
9040 parsers expect JSON objects. The new output mode
9041 "json-pretty" has been added to provide similar output, but
9042 neatly aligned for readability by humans.
9044 * We dropped all explicit sync() invocations in the shutdown
9045 code. The kernel does this implicitly anyway in the kernel
9046 reboot() syscall. halt(8)'s -n option is now a compatibility
9049 * We now support virtualized reboot() in containers, as
9050 supported by newer kernels. We will fall back to exit() if
9051 CAP_SYS_REBOOT is not available to the container. Also,
9052 nspawn makes use of this now and will actually reboot the
9053 container if the containerized OS asks for that.
9055 * journalctl will only show local log output by default
9056 now. Use --merge (-m) to show remote log output, too.
9058 * libsystemd-journal gained the new sd_journal_get_usage()
9059 call to determine the current disk usage of all journal
9060 files. This is exposed in the new "journalctl --disk-usage"
9063 * journald gained a new configuration setting SplitMode= in
9064 journald.conf which may be used to control how user journals
9065 are split off. See journald.conf(5) for details.
9067 * A new condition type ConditionFileNotEmpty= has been added.
9069 * tmpfiles' "w" lines now support file globbing, to write
9070 multiple files at once.
9072 * We added Python bindings for the journal submission
9073 APIs. More Python APIs for a number of selected APIs will
9074 likely follow. Note that we intend to add native bindings
9075 only for the Python language, as we consider it common
9076 enough to deserve bindings shipped within systemd. There are
9077 various projects outside of systemd that provide bindings
9078 for languages such as PHP or Lua.
9080 * Many conditions will now resolve specifiers such as %i. In
9081 addition, PathChanged= and related directives of .path units
9082 now support specifiers as well.
9084 * There's now a new RPM macro definition for the system preset
9087 * journald will now warn if it ca not forward a message to the
9088 syslog daemon because its socket is full.
9090 * timedated will no longer write or process /etc/timezone,
9091 except on Debian. As we do not support late mounted /usr
9092 anymore /etc/localtime always being a symlink is now safe,
9093 and hence the information in /etc/timezone is not necessary
9096 * logind will now always reserve one VT for a text getty (VT6
9097 by default). Previously if more than 6 X sessions where
9098 started they took up all the VTs with auto-spawned gettys,
9099 so that no text gettys were available anymore.
9101 * udev will now automatically inform the btrfs kernel logic
9102 about btrfs RAID components showing up. This should make
9103 simple hotplug based btrfs RAID assembly work.
9105 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
9106 (but not for its children which will stay at the kernel
9107 default). This should allow setups with a lot more listening
9110 * systemd will now always pass the configured timezone to the
9111 kernel at boot. timedated will do the same when the timezone
9114 * logind's inhibition logic has been updated. By default,
9115 logind will now handle the lid switch, the power and sleep
9116 keys all the time, even in graphical sessions. If DEs want
9117 to handle these events on their own they should take the new
9118 handle-power-key, handle-sleep-key and handle-lid-switch
9119 inhibitors during their runtime. A simple way to achieve
9120 that is to invoke the DE wrapped in an invocation of:
9122 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
9124 * Access to unit operations is now checked via SELinux taking
9125 the unit file label and client process label into account.
9127 * systemd will now notify the administrator in the journal
9128 when he over-mounts a non-empty directory.
9130 * There are new specifiers that are resolved in unit files,
9131 for the host name (%H), the machine ID (%m) and the boot ID
9134 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
9135 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
9136 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
9137 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
9138 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
9139 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
9140 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
9144 * Support for reading structured kernel messages from
9145 /dev/kmsg has now been added and is enabled by default.
9147 * Support for reading kernel messages from /proc/kmsg has now
9148 been removed. If you want kernel messages in the journal
9149 make sure to run a recent kernel (>= 3.5) that supports
9150 reading structured messages from /dev/kmsg (see
9151 above). /proc/kmsg is now exclusive property of classic
9152 syslog daemons again.
9154 * The libudev API gained the new
9155 udev_device_new_from_device_id() call.
9157 * The logic for file system namespace (ReadOnlyDirectory=,
9158 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
9159 require pivot_root() anymore. This means fewer temporary
9160 directories are created below /tmp for this feature.
9162 * nspawn containers will now see and receive all submounts
9163 made on the host OS below the root file system of the
9166 * Forward Secure Sealing is now supported for Journal files,
9167 which provide cryptographical sealing of journal files so
9168 that attackers cannot alter log history anymore without this
9169 being detectable. Lennart will soon post a blog story about
9170 this explaining it in more detail.
9172 * There are two new service settings RestartPreventExitStatus=
9173 and SuccessExitStatus= which allow configuration of exit
9174 status (exit code or signal) which will be excepted from the
9175 restart logic, resp. consider successful.
9177 * journalctl gained the new --verify switch that can be used
9178 to check the integrity of the structure of journal files and
9179 (if Forward Secure Sealing is enabled) the contents of
9182 * nspawn containers will now be run with /dev/stdin, /dev/fd/
9183 and similar symlinks pre-created. This makes running shells
9184 as container init process a lot more fun.
9186 * The fstab support can now handle PARTUUID= and PARTLABEL=
9189 * A new ConditionHost= condition has been added to match
9190 against the hostname (with globs) and machine ID. This is
9191 useful for clusters where a single OS image is used to
9192 provision a large number of hosts which shall run slightly
9193 different sets of services.
9195 * Services which hit the restart limit will now be placed in a
9198 Contributions from: Bertram Poettering, Dave Reisner, Huang
9199 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
9200 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
9204 * When running in --user mode systemd will now become a
9205 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
9206 tree a lot more organized.
9208 * A new PartOf= unit dependency type has been introduced that
9209 may be used to group services in a natural way.
9211 * "systemctl enable" may now be used to enable instances of
9214 * journalctl now prints error log levels in red, and
9215 warning/notice log levels in bright white. It also supports
9216 filtering by log level now.
9218 * cgtop gained a new -n switch (similar to top), to configure
9219 the maximum number of iterations to run for. It also gained
9220 -b, to run in batch mode (accepting no input).
9222 * The suffix ".service" may now be omitted on most systemctl
9223 command lines involving service unit names.
9225 * There's a new bus call in logind to lock all sessions, as
9226 well as a loginctl verb for it "lock-sessions".
9228 * libsystemd-logind.so gained a new call sd_journal_perror()
9229 that works similar to libc perror() but logs to the journal
9230 and encodes structured information about the error number.
9232 * /etc/crypttab entries now understand the new keyfile-size=
9235 * shutdown(8) now can send a (configurable) wall message when
9236 a shutdown is cancelled.
9238 * The mount propagation mode for the root file system will now
9239 default to "shared", which is useful to make containers work
9240 nicely out-of-the-box so that they receive new mounts from
9241 the host. This can be undone locally by running "mount
9242 --make-rprivate /" if needed.
9244 * The prefdm.service file has been removed. Distributions
9245 should maintain this unit downstream if they intend to keep
9246 it around. However, we recommend writing normal unit files
9247 for display managers instead.
9249 * Since systemd is a crucial part of the OS we will now
9250 default to a number of compiler switches that improve
9251 security (hardening) such as read-only relocations, stack
9252 protection, and suchlike.
9254 * The TimeoutSec= setting for services is now split into
9255 TimeoutStartSec= and TimeoutStopSec= to allow configuration
9256 of individual time outs for the start and the stop phase of
9259 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
9260 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
9261 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
9262 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
9263 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
9264 Gundersen, Zbigniew Jędrzejewski-Szmek
9268 * The journal and id128 C APIs are now fully documented as man
9271 * Extra safety checks have been added when transitioning from
9272 the initial RAM disk to the main system to avoid accidental
9275 * /etc/crypttab entries now understand the new keyfile-offset=
9278 * systemctl -t can now be used to filter by unit load state.
9280 * The journal C API gained the new sd_journal_wait() call to
9281 make writing synchronous journal clients easier.
9283 * journalctl gained the new -D switch to show journals from a
9286 * journalctl now displays a special marker between log
9287 messages of two different boots.
9289 * The journal is now explicitly flushed to /var via a service
9290 systemd-journal-flush.service, rather than implicitly simply
9291 by seeing /var/log/journal to be writable.
9293 * journalctl (and the journal C APIs) can now match for much
9294 more complex expressions, with alternatives and
9297 * When transitioning from the initial RAM disk to the main
9298 system we will now kill all processes in a killing spree to
9299 ensure no processes stay around by accident.
9301 * Three new specifiers may be used in unit files: %u, %h, %s
9302 resolve to the user name, user home directory resp. user
9303 shell. This is useful for running systemd user instances.
9305 * We now automatically rotate journal files if their data
9306 object hash table gets a fill level > 75%. We also size the
9307 hash table based on the configured maximum file size. This
9308 together should lower hash collisions drastically and thus
9309 speed things up a bit.
9311 * journalctl gained the new "--header" switch to introspect
9312 header data of journal files.
9314 * A new setting SystemCallFilters= has been added to services
9315 which may be used to apply blacklists or whitelists to
9316 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
9318 * nspawn gained a new --link-journal= switch (and quicker: -j)
9319 to link the container journal with the host. This makes it
9320 very easy to centralize log viewing on the host for all
9321 guests while still keeping the journal files separated.
9323 * Many bugfixes and optimizations
9325 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
9326 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
9327 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
9332 * Several tools now understand kernel command line arguments,
9333 which are only read when run in an initial RAM disk. They
9334 usually follow closely their normal counterparts, but are
9337 * There's a new tool to analyze the readahead files that are
9338 automatically generated at boot. Use:
9340 /usr/lib/systemd/systemd-readahead analyze /.readahead
9342 * We now provide an early debug shell on tty9 if this enabled. Use:
9344 systemctl enable debug-shell.service
9346 * All plymouth related units have been moved into the Plymouth
9347 package. Please make sure to upgrade your Plymouth version
9350 * systemd-tmpfiles now supports getting passed the basename of
9351 a configuration file only, in which case it will look for it
9352 in all appropriate directories automatically.
9354 * udevadm info now takes a /dev or /sys path as argument, and
9355 does the right thing. Example:
9357 udevadm info /dev/sda
9358 udevadm info /sys/class/block/sda
9360 * systemctl now prints a warning if a unit is stopped but a
9361 unit that might trigger it continues to run. Example: a
9362 service is stopped but the socket that activates it is left
9365 * "systemctl status" will now mention if the log output was
9366 shortened due to rotation since a service has been started.
9368 * The journal API now exposes functions to determine the
9369 "cutoff" times due to rotation.
9371 * journald now understands SIGUSR1 and SIGUSR2 for triggering
9372 immediately flushing of runtime logs to /var if possible,
9373 resp. for triggering immediate rotation of the journal
9376 * It is now considered an error if a service is attempted to
9377 be stopped that is not loaded.
9379 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
9381 * systemd-analyze now supports Python 3
9383 * tmpfiles now supports cleaning up directories via aging
9384 where the first level dirs are always kept around but
9385 directories beneath it automatically aged. This is enabled
9386 by prefixing the age field with '~'.
9388 * Seat objects now expose CanGraphical, CanTTY properties
9389 which is required to deal with very fast bootups where the
9390 display manager might be running before the graphics drivers
9391 completed initialization.
9393 * Seat objects now expose a State property.
9395 * We now include RPM macros for service enabling/disabling
9396 based on the preset logic. We recommend RPM based
9397 distributions to make use of these macros if possible. This
9398 makes it simpler to reuse RPM spec files across
9401 * We now make sure that the collected systemd unit name is
9402 always valid when services log to the journal via
9405 * There's a new man page kernel-command-line(7) detailing all
9406 command line options we understand.
9408 * The fstab generator may now be disabled at boot by passing
9409 fstab=0 on the kernel command line.
9411 * A new kernel command line option modules-load= is now understood
9412 to load a specific kernel module statically, early at boot.
9414 * Unit names specified on the systemctl command line are now
9415 automatically escaped as needed. Also, if file system or
9416 device paths are specified they are automatically turned
9417 into the appropriate mount or device unit names. Example:
9419 systemctl status /home
9420 systemctl status /dev/sda
9422 * The SysVConsole= configuration option has been removed from
9423 system.conf parsing.
9425 * The SysV search path is no longer exported on the D-Bus
9428 * The Names= option has been removed from unit file parsing.
9430 * There's a new man page bootup(7) detailing the boot process.
9432 * Every unit and every generator we ship with systemd now
9433 comes with full documentation. The self-explanatory boot is
9436 * A couple of services gained "systemd-" prefixes in their
9437 name if they wrap systemd code, rather than only external
9438 code. Among them fsck@.service which is now
9439 systemd-fsck@.service.
9441 * The HaveWatchdog property has been removed from the D-Bus
9444 * systemd.confirm_spawn= on the kernel command line should now
9447 * There's a new man page crypttab(5) which details all options
9448 we actually understand.
9450 * systemd-nspawn gained a new --capability= switch to pass
9451 additional capabilities to the container.
9453 * timedated will now read known NTP implementation unit names
9454 from /usr/lib/systemd/ntp-units.d/*.list,
9455 systemd-timedated-ntp.target has been removed.
9457 * journalctl gained a new switch "-b" that lists log data of
9458 the current boot only.
9460 * The notify socket is in the abstract namespace again, in
9461 order to support daemons which chroot() at start-up.
9463 * There is a new Storage= configuration option for journald
9464 which allows configuration of where log data should go. This
9465 also provides a way to disable journal logging entirely, so
9466 that data collected is only forwarded to the console, the
9467 kernel log buffer or another syslog implementation.
9469 * Many bugfixes and optimizations
9471 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
9472 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
9473 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
9474 Shawn Landden, Tom Gundersen
9478 * "systemctl help <unit>" now shows the man page if one is
9481 * Several new man pages have been added.
9483 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
9484 MaxLevelConsole= can now be specified in
9485 journald.conf. These options allow reducing the amount of
9486 data stored on disk or forwarded by the log level.
9488 * TimerSlackNSec= can now be specified in system.conf for
9489 PID1. This allows system-wide power savings.
9491 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
9492 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
9497 * logind is now capable of (optionally) handling power and
9498 sleep keys as well as the lid switch.
9500 * journalctl now understands the syntax "journalctl
9501 /usr/bin/avahi-daemon" to get all log output of a specific
9504 * CapabilityBoundingSet= in system.conf now also influences
9505 the capability bound set of usermode helpers of the kernel.
9507 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
9508 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
9509 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
9510 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
9514 * Note that we skipped 139 releases here in order to set the
9515 new version to something that is greater than both udev's
9516 and systemd's most recent version number.
9518 * udev: all udev sources are merged into the systemd source tree now.
9519 All future udev development will happen in the systemd tree. It
9520 is still fully supported to use the udev daemon and tools without
9521 systemd running, like in initramfs or other init systems. Building
9522 udev though, will require the *build* of the systemd tree, but
9523 udev can be properly *run* without systemd.
9525 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
9526 should be used to create dead device nodes as workarounds for broken
9529 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
9530 no longer supported. udev_monitor_new_from_netlink() needs to be
9531 used to subscribe to events.
9533 * udev: when udevd is started by systemd, processes which are left
9534 behind by forking them off of udev rules, are unconditionally cleaned
9535 up and killed now after the event handling has finished. Services or
9536 daemons must be started as systemd services. Services can be
9537 pulled-in by udev to get started, but they can no longer be directly
9538 forked by udev rules.
9540 * udev: the daemon binary is called systemd-udevd now and installed
9541 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
9542 to adapt to that, create symlink, or rename the binary after building
9545 * libudev no longer provides these symbols:
9546 udev_monitor_from_socket()
9547 udev_queue_get_failed_list_entry()
9548 udev_get_{dev,sys,run}_path()
9549 The versions number was bumped and symbol versioning introduced.
9551 * systemd-loginctl and systemd-journalctl have been renamed
9552 to loginctl and journalctl to match systemctl.
9554 * The config files: /etc/systemd/systemd-logind.conf and
9555 /etc/systemd/systemd-journald.conf have been renamed to
9556 logind.conf and journald.conf. Package updates should rename
9557 the files to the new names on upgrade.
9559 * For almost all files the license is now LGPL2.1+, changed
9560 from the previous GPL2.0+. Exceptions are some minor stuff
9561 of udev (which will be changed to LGPL2.1 eventually, too),
9562 and the MIT licensed sd-daemon.[ch] library that is suitable
9563 to be used as drop-in files.
9565 * systemd and logind now handle system sleep states, in
9566 particular suspending and hibernating.
9568 * logind now implements a sleep/shutdown/idle inhibiting logic
9569 suitable for a variety of uses. Soonishly Lennart will blog
9570 about this in more detail.
9572 * var-run.mount and var-lock.mount are no longer provided
9573 (which previously bind mounted these directories to their new
9574 places). Distributions which have not converted these
9575 directories to symlinks should consider stealing these files
9576 from git history and add them downstream.
9578 * We introduced the Documentation= field for units and added
9579 this to all our shipped units. This is useful to make it
9580 easier to explore the boot and the purpose of the various
9583 * All smaller setup units (such as
9584 systemd-vconsole-setup.service) now detect properly if they
9585 are run in a container and are skipped when
9586 appropriate. This guarantees an entirely noise-free boot in
9587 Linux container environments such as systemd-nspawn.
9589 * A framework for implementing offline system updates is now
9590 integrated, for details see:
9591 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
9593 * A new service type Type=idle is available now which helps us
9594 avoiding ugly interleaving of getty output and boot status
9597 * There's now a system-wide CapabilityBoundingSet= option to
9598 globally reduce the set of capabilities for the
9599 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
9600 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
9601 even CAP_NET_ADMIN system-wide for secure systems.
9603 * There are now system-wide DefaultLimitXXX= options to
9604 globally change the defaults of the various resource limits
9605 for all units started by PID 1.
9607 * Harald Hoyer's systemd test suite has been integrated into
9608 systemd which allows easy testing of systemd builds in qemu
9609 and nspawn. (This is really awesome! Ask us for details!)
9611 * The fstab parser is now implemented as generator, not inside
9614 * systemctl will now warn you if .mount units generated from
9615 /etc/fstab are out of date due to changes in fstab that
9616 have not been read by systemd yet.
9618 * systemd is now suitable for usage in initrds. Dracut has
9619 already been updated to make use of this. With this in place
9620 initrds get a slight bit faster but primarily are much
9621 easier to introspect and debug since "systemctl status" in
9622 the host system can be used to introspect initrd services,
9623 and the journal from the initrd is kept around too.
9625 * systemd-delta has been added, a tool to explore differences
9626 between user/admin configuration and vendor defaults.
9628 * PrivateTmp= now affects both /tmp and /var/tmp.
9630 * Boot time status messages are now much prettier and feature
9631 proper english language. Booting up systemd has never been
9634 * Read-ahead pack files now include the inode number of all
9635 files to pre-cache. When the inode changes the pre-caching
9636 is not attempted. This should be nicer to deal with updated
9637 packages which might result in changes of read-ahead
9640 * We now temporaritly lower the kernel's read_ahead_kb variable
9641 when collecting read-ahead data to ensure the kernel's
9642 built-in read-ahead does not add noise to our measurements
9643 of necessary blocks to pre-cache.
9645 * There's now RequiresMountsFor= to add automatic dependencies
9646 for all mounts necessary for a specific file system path.
9648 * MountAuto= and SwapAuto= have been removed from
9649 system.conf. Mounting file systems at boot has to take place
9652 * nspawn now learned a new switch --uuid= to set the machine
9653 ID on the command line.
9655 * nspawn now learned the -b switch to automatically search
9658 * vt102 is now the default TERM for serial TTYs, upgraded from
9661 * systemd-logind now works on VT-less systems.
9663 * The build tree has been reorganized. The individual
9664 components now have directories of their own.
9666 * A new condition type ConditionPathIsReadWrite= is now available.
9668 * nspawn learned the new -C switch to create cgroups for the
9669 container in other hierarchies.
9671 * We now have support for hardware watchdogs, configurable in
9674 * The scheduled shutdown logic now has a public API.
9676 * We now mount /tmp as tmpfs by default, but this can be
9677 masked and /etc/fstab can override it.
9679 * Since udisks does not make use of /media anymore we are not
9680 mounting a tmpfs on it anymore.
9682 * journalctl gained a new --local switch to only interleave
9683 locally generated journal files.
9685 * We can now load the IMA policy at boot automatically.
9687 * The GTK tools have been split off into a systemd-ui.
9689 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
9690 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
9691 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
9692 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
9693 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
9694 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
9695 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
9696 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
9697 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
9702 * This is mostly a bugfix release
9704 * Support optional initialization of the machine ID from the
9705 KVM or container configured UUID.
9707 * Support immediate reboots with "systemctl reboot -ff"
9709 * Show /etc/os-release data in systemd-analyze output
9711 * Many bugfixes for the journal, including endianness fixes and
9712 ensuring that disk space enforcement works
9714 * sd-login.h is C++ compatible again
9716 * Extend the /etc/os-release format on request of the Debian
9719 * We now refuse non-UTF8 strings used in various configuration
9720 and unit files. This is done to ensure we do not pass invalid
9721 data over D-Bus or expose it elsewhere.
9723 * Register Mimo USB Screens as suitable for automatic seat
9726 * Read SELinux client context from journal clients in a race
9729 * Reorder configuration file lookup order. /etc now always
9730 overrides /run in order to allow the administrator to always
9731 and unconditionally override vendor-supplied or
9732 automatically generated data.
9734 * The various user visible bits of the journal now have man
9735 pages. We still lack man pages for the journal API calls
9738 * We now ship all man pages in HTML format again in the
9741 Contributions from: Dave Reisner, Dirk Eibach, Frederic
9742 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
9743 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
9748 * This is mostly a bugfix release
9750 * systems lacking /etc/os-release are no longer supported.
9752 * Various functionality updates to libsystemd-login.so
9754 * Track class of PAM logins to distinguish greeters from
9757 Contributions from: Kay Sievers, Lennart Poettering, Michael
9762 * This is an important bugfix release for v41.
9764 * Building man pages is now optional which should be useful
9765 for those building systemd from git but unwilling to install
9768 * Watchdog support for supervising services is now usable. In
9769 a future release support for hardware watchdogs
9770 (i.e. /dev/watchdog) will be added building on this.
9772 * Service start rate limiting is now configurable and can be
9773 turned off per service. When a start rate limit is hit a
9774 reboot can automatically be triggered.
9776 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
9778 Contributions from: Benjamin Franzke, Bill Nottingham,
9779 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
9780 Schmidt, Michał Górny, Piotr Drąg
9784 * The systemd binary is installed /usr/lib/systemd/systemd now;
9785 An existing /sbin/init symlink needs to be adapted with the
9788 * The code that loads kernel modules has been ported to invoke
9789 libkmod directly, instead of modprobe. This means we do not
9790 support systems with module-init-tools anymore.
9792 * Watchdog support is now already useful, but still not
9795 * A new kernel command line option systemd.setenv= is
9796 understood to set system wide environment variables
9797 dynamically at boot.
9799 * We now limit the set of capabilities of systemd-journald.
9801 * We now set SIGPIPE to ignore by default, since it only is
9802 useful in shell pipelines, and has little use in general
9803 code. This can be disabled with IgnoreSIPIPE=no in unit
9806 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
9807 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
9812 * This is mostly a bugfix release
9814 * We now expose the reason why a service failed in the
9815 "Result" D-Bus property.
9817 * Rudimentary service watchdog support (will be completed over
9818 the next few releases.)
9820 * When systemd forks off in order execute some service we will
9821 now immediately changes its argv[0] to reflect which process
9822 it will execute. This is useful to minimize the time window
9823 with a generic argv[0], which makes bootcharts more useful
9825 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
9826 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
9827 Mike Kazantsev, Ray Strode
9831 * This is mostly a test release, but incorporates many
9834 * New systemd-cgtop tool to show control groups by their
9837 * Linking against libacl for ACLs is optional again. If
9838 disabled, support tracking device access for active logins
9839 goes becomes unavailable, and so does access to the user
9840 journals by the respective users.
9842 * If a group "adm" exists, journal files are automatically
9843 owned by them, thus allow members of this group full access
9844 to the system journal as well as all user journals.
9846 * The journal now stores the SELinux context of the logging
9847 client for all entries.
9849 * Add C++ inclusion guards to all public headers
9851 * New output mode "cat" in the journal to print only text
9852 messages, without any meta data like date or time.
9854 * Include tiny X server wrapper as a temporary stop-gap to
9855 teach XOrg udev display enumeration. This is used by display
9856 managers such as gdm, and will go away as soon as XOrg
9857 learned native udev hotplugging for display devices.
9859 * Add new systemd-cat tool for executing arbitrary programs
9860 with STDERR/STDOUT connected to the journal. Can also act as
9861 BSD logger replacement, and does so by default.
9863 * Optionally store all locally generated coredumps in the
9864 journal along with meta data.
9866 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
9867 writing short strings to files (for usage for /sys), and for
9868 creating symlinks, character and block device nodes.
9870 * New unit file option ControlGroupPersistent= to make cgroups
9871 persistent, following the mechanisms outlined in
9872 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
9874 * Support multiple local RTCs in a sane way
9876 * No longer monopolize IO when replaying readahead data on
9877 rotating disks, since we might starve non-file-system IO to
9878 death, since fanotify() will not see accesses done by blkid,
9881 * Do not show kernel threads in systemd-cgls anymore, unless
9882 requested with new -k switch.
9884 Contributions from: Dan Horák, Kay Sievers, Lennart
9885 Poettering, Michal Schmidt
9889 * This is mostly a test release, but incorporates many
9892 * The git repository moved to:
9893 git://anongit.freedesktop.org/systemd/systemd
9894 ssh://git.freedesktop.org/git/systemd/systemd
9896 * First release with the journal
9897 http://0pointer.de/blog/projects/the-journal.html
9899 * The journal replaces both systemd-kmsg-syslogd and
9900 systemd-stdout-bridge.
9902 * New sd_pid_get_unit() API call in libsystemd-logind
9904 * Many systemadm clean-ups
9906 * Introduce remote-fs-pre.target which is ordered before all
9907 remote mounts and may be used to start services before all
9910 * Added Mageia support
9912 * Add bash completion for systemd-loginctl
9914 * Actively monitor PID file creation for daemons which exit in
9915 the parent process before having finished writing the PID
9916 file in the daemon process. Daemons which do this need to be
9917 fixed (i.e. PID file creation must have finished before the
9918 parent exits), but we now react a bit more gracefully to them.
9920 * Add colourful boot output, mimicking the well-known output
9921 of existing distributions.
9923 * New option PassCredentials= for socket units, for
9924 compatibility with a recent kernel ABI breakage.
9926 * /etc/rc.local is now hooked in via a generator binary, and
9927 thus will no longer act as synchronization point during
9930 * systemctl list-unit-files now supports --root=.
9932 * systemd-tmpfiles now understands two new commands: z, Z for
9933 relabelling files according to the SELinux database. This is
9934 useful to apply SELinux labels to specific files in /sys,
9937 * Output of SysV services is now forwarded to both the console
9938 and the journal by default, not only just the console.
9940 * New man pages for all APIs from libsystemd-login.
9942 * The build tree got reorganized and the build system is a
9943 lot more modular allowing embedded setups to specifically
9944 select the components of systemd they are interested in.
9946 * Support for Linux systems lacking the kernel VT subsystem is
9949 * configure's --with-rootdir= got renamed to
9950 --with-rootprefix= to follow the naming used by udev and
9953 * Unless specified otherwise we will now install to /usr instead
9954 of /usr/local by default.
9956 * Processes with '@' in argv[0][0] are now excluded from the
9957 final shut-down killing spree, following the logic explained
9959 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
9961 * All processes remaining in a service cgroup when we enter
9962 the START or START_PRE states are now killed with
9963 SIGKILL. That means it is no longer possible to spawn
9964 background processes from ExecStart= lines (which was never
9965 supported anyway, and bad style).
9967 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
9968 reloading of units together.
9970 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
9971 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
9972 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
9973 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
9974 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek