1 systemd System and Service Manager
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 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
283 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
284 IPv6 configuration (static or DHCPv6) was found for a given
285 interface. It will not touch the sysctl otherwise.
287 * The order of entries is $PATH used by the user manager instance was
288 changed to put bin/ entries before the corresponding sbin/ entries.
289 It is recommended to not rely on this order, and only ever have one
290 binary with a given name in the system paths under /usr.
292 * A new tool systemd-network-generator has been added that may generate
293 .network, .netdev and .link files from IP configuration specified on
294 the kernel command line in the format used by Dracut.
296 * The CriticalConnection= setting in .network files is now deprecated,
297 and replaced by a new KeepConfiguration= setting which allows more
298 detailed configuration of the IP configuration to keep in place.
300 * systemd-analyze gained a few new verbs:
302 - "systemd-analyze timestamp" parses and converts timestamps. This is
303 similar to the existing "systemd-analyze calendar" command which
304 does the same for recurring calendar events.
306 - "systemd-analyze timespan" parses and converts timespans (i.e.
307 durations as opposed to points in time).
309 - "systemd-analyze condition" will parse and test ConditionXYZ=
312 - "systemd-analyze exit-status" will parse and convert exit status
313 codes to their names and back.
315 - "systemd-analyze unit-files" will print a list of all unit
316 file paths and unit aliases.
318 * SuccessExitStatus=, RestartPreventExitStatus=, and
319 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
320 is equivalent to "65"). Those exit status name mappings may be
321 displayed with the sytemd-analyze exit-status verb describe above.
323 * systemd-logind now exposes a per-session SetBrightness() bus call,
324 which may be used to securely change the brightness of a kernel
325 brightness device, if it belongs to the session's seat. By using this
326 call unprivileged clients can make changes to "backlight" and "leds"
327 devices securely with strict requirements on session membership.
328 Desktop environments may use this to generically make brightness
329 changes to such devices without shipping private SUID binaries or
330 udev rules for that purpose.
332 * "udevadm info" gained a --wait-for-initialization switch to wait for
333 a device to be initialized.
335 * systemd-hibernate-resume-generator will now look for resumeflags= on
336 the kernel command line, which is similar to rootflags= and may be
337 used to configure device timeout for the hibernation device.
339 * sd-event learnt a new API call sd_event_source_disable_unref() for
340 disabling and unref'ing an event source in a single function. A
341 related call sd_event_source_disable_unrefp() has been added for use
342 with gcc's cleanup extension.
344 * The sd-id128.h public API gained a new definition
345 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
348 * "busctl introspect" gained a new switch --xml-interface for dumping
349 XML introspection data unmodified.
351 * PID 1 may now show the unit name instead of the unit description
352 string in its status output during boot. This may be configured in
353 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
354 kernel command line option systemd.status_unit_format=.
356 * PID 1 now understands a new option KExecWatchdogSec= in
357 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
358 Previously watchdog functionality was only available for regular
359 reboots. The new setting defaults to off, because we don't know in
360 the general case if the watchdog will be reset after kexec (some
361 drivers do reset it, but not all), and the new userspace might not be
362 configured to handle the watchdog.
364 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
365 RebootWatchdogSec= to more clearly communicate what it is about. The
366 old name is still accepted for compatibility.
368 * The systemd.debug_shell kernel command line option now optionally
369 takes a tty name to spawn the debug shell on, which allows a
370 different tty to be selected than the built-in default.
372 * Service units gained a new ExecCondition= setting which will run
373 before ExecStartPre= and either continue execution of the unit (for
374 clean exit codes), stop execution without marking the unit failed
375 (for exit codes 1 through 254), or stop execution and fail the unit
376 (for exit code 255 or abnormal termination).
378 * A new service systemd-pstore.service has been added that pulls data
379 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
382 * timedatectl gained new verbs for configuring per-interface NTP
383 service configuration for systemd-timesyncd.
385 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
386 2019. (You can set non-UTF-8 locales though, if you know their name.)
388 * If variable assignments in sysctl.d/ files are prefixed with "-" any
389 failures to apply them are now ignored.
391 * systemd-random-seed.service now optionally credits entropy when
392 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
393 true for the service to enable this behaviour, but please consult the
394 documentation first, since this comes with a couple of caveats.
396 * systemd-random-seed.service is now a synchronization point for full
397 initialization of the kernel's entropy pool. Services that require
398 /dev/urandom to be correctly initialized should be ordered after this
401 * The systemd-boot boot loader has been updated to optionally maintain
402 a random seed file in the EFI System Partition (ESP). During the boot
403 phase, this random seed is read and updated with a new seed
404 cryptographically derived from it. Another derived seed is passed to
405 the OS. The latter seed is then credited to the kernel's entropy pool
406 very early during userspace initialization (from PID 1). This allows
407 systems to boot up with a fully initialized kernel entropy pool from
408 earliest boot on, and thus entirely removes all entropy pool
409 initialization delays from systems using systemd-boot. Special care
410 is taken to ensure different seeds are derived on system images
411 replicated to multiple systems. "bootctl status" will show whether
412 a seed was received from the boot loader.
414 * bootctl gained two new verbs:
416 - "bootctl random-seed" will generate the file in ESP and an EFI
417 variable to allow a random seed to be passed to the OS as described
420 - "bootctl is-installed" checks whether systemd-boot is currently
423 * bootctl will warn if it detects that boot entries are misconfigured
424 (for example if the kernel image was removed without purging the
427 * A new document has been added describing systemd's use and support
428 for the kernel's entropy pool subsystem:
430 https://systemd.io/RANDOM_SEEDS
432 * When the system is hibernated the swap device to write the
433 hibernation image to is now automatically picked from all available
434 swap devices, preferring the swap device with the highest configured
435 priority over all others, and picking the device with the most free
436 space if there are multiple devices with the highest priority.
438 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
439 option that permits selecting the timout how long to wait for a
440 device with an encryption key before asking for the password.
442 * IOWeight= has learnt to properly set the IO weight when using the
443 BFQ scheduler officially found in kernels 5.0+.
445 * A new mailing list has been created for reporting of security issues:
446 systemd-security@redhat.com. For mode details, see
447 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
449 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
450 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
451 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
452 Chiu, Chris Down, Christian Kellner, Clinton Roy, Connor Reeder, Daniel
453 Black, Daniele Medri, Dan Streetman, Dave Reisner, Dave Ross, David
454 Art, David Tardon, Debarshi Ray, Dimitri John Ledkov, Dominick Grift,
455 Donald Buczek, Douglas Christman, Eric DeVolder, EtherGraf, Evgeny
456 Vereshchagin, Feldwor, Felix Riemann, Florian Dollinger, Francesco
457 Pennica, Franck Bui, Frantisek Sumsal, Franz Pletz, frederik, Hans
458 de Goede, Iago López Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer,
459 Jack, Jakob Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan
460 Pokorný, Jan Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller,
461 Jérémy Rosen, Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann
462 B. Guðmundsson, Johannes Christ, Johannes Schmitz, Jonathan Rouleau,
463 Jorge Niedbalski, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
464 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
465 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
466 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
467 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
468 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
469 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
470 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Roberto
471 Santalla, Ronan Pigott, root, RussianNeuroMancer, Sebastian Jennen,
472 shinygold, Shreyas Behera, Simon Schricker, Susant Sahani, Thadeu Lima
473 de Souza Cascardo, Theo Ouzhinski, Thiebaud Weksteen, Thomas Haller,
474 Thomas Weißschuh, Tomas Mraz, Tommi Rantala, Topi Miettinen, VD-Lycos,
475 ven, Wieland Hoffmann, William A. Kennington III, William Wold, Xi
476 Ruoyao, Yuri Chornoivan, Yu Watanabe, Your Name, Zach Smith, Zbigniew
477 Jędrzejewski-Szmek, Zhang Xianwei
479 – Camerino, 2019-09-03
483 * In .link files, MACAddressPolicy=persistent (the default) is changed
484 to cover more devices. For devices like bridges, tun, tap, bond, and
485 similar interfaces that do not have other identifying information,
486 the interface name is used as the basis for persistent seed for MAC
487 and IPv4LL addresses. The way that devices that were handled
488 previously is not changed, and this change is about covering more
489 devices then previously by the "persistent" policy.
491 MACAddressPolicy=random may be used to force randomized MACs and
492 IPv4LL addresses for a device if desired.
494 Hint: the log output from udev (at debug level) was enhanced to
495 clarify what policy is followed and which attributes are used.
496 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
497 may be used to view this.
499 Hint: if a bridge interface is created without any slaves, and gains
500 a slave later, then now the bridge does not inherit slave's MAC.
501 To inherit slave's MAC, for example, create the following file:
503 # /etc/systemd/network/98-bridge-inherit-mac.link
508 MACAddressPolicy=none
511 * The .device units generated by systemd-fstab-generator and other
512 generators do not automatically pull in the corresponding .mount unit
513 as a Wants= dependency. This means that simply plugging in the device
514 will not cause the mount unit to be started automatically. But please
515 note that the mount unit may be started for other reasons, in
516 particular if it is part of local-fs.target, and any unit which
517 (transitively) depends on local-fs.target is started.
519 * networkctl list/status/lldp now accept globbing wildcards for network
520 interface names to match against all existing interfaces.
522 * The $PIDFILE environment variable is set to point the absolute path
523 configured with PIDFile= for processes of that service.
525 * The fallback DNS server list was augmented with Cloudflare public DNS
526 servers. Use `-Ddns-servers=` to set a different fallback.
528 * A new special target usb-gadget.target will be started automatically
529 when a USB Device Controller is detected (which means that the system
530 is a USB peripheral).
532 * A new unit setting CPUQuotaPeriodSec= assigns the time period
533 relatively to which the CPU time quota specified by CPUQuota= is
536 * A new unit setting ProtectHostname= may be used to prevent services
537 from modifying hostname information (even if they otherwise would
538 have privileges to do so).
540 * A new unit setting NetworkNamespacePath= may be used to specify a
541 namespace for service or socket units through a path referring to a
542 Linux network namespace pseudo-file.
544 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
545 have an effect on .socket units: when used the listening socket is
546 created within the configured network namespace instead of the host
549 * ExecStart= command lines in unit files may now be prefixed with ':'
550 in which case environment variable substitution is
551 disabled. (Supported for the other ExecXYZ= settings, too.)
553 * .timer units gained two new boolean settings OnClockChange= and
554 OnTimezoneChange= which may be used to also trigger a unit when the
555 system clock is changed or the local timezone is
556 modified. systemd-run has been updated to make these options easily
557 accessible from the command line for transient timers.
559 * Two new conditions for units have been added: ConditionMemory= may be
560 used to conditionalize a unit based on installed system
561 RAM. ConditionCPUs= may be used to conditionalize a unit based on
564 * The @default system call filter group understood by SystemCallFilter=
565 has been updated to include the new rseq() system call introduced in
568 * A new time-set.target has been added that indicates that the system
569 time has been set from a local source (possibly imprecise). The
570 existing time-sync.target is stronger and indicates that the time has
571 been synchronized with a precise external source. Services where
572 approximate time is sufficient should use the new target.
574 * "systemctl start" (and related commands) learnt a new
575 --show-transaction option. If specified brief information about all
576 jobs queued because of the requested operation is shown.
578 * systemd-networkd recognizes a new operation state 'enslaved', used
579 (instead of 'degraded' or 'carrier') for interfaces which form a
580 bridge, bond, or similar, and an new 'degraded-carrier' operational
581 state used for the bond or bridge master interface when one of the
582 enslaved devices is not operational.
584 * .network files learnt the new IgnoreCarrierLoss= option for leaving
585 networks configured even if the carrier is lost.
587 * The RequiredForOnline= setting in .network files may now specify a
588 minimum operational state required for the interface to be considered
589 "online" by systemd-networkd-wait-online. Related to this
590 systemd-networkd-wait-online gained a new option --operational-state=
591 to configure the same, and its --interface= option was updated to
592 optionally also take an operational state specific for an interface.
594 * systemd-networkd-wait-online gained a new setting --any for waiting
595 for only one of the requested interfaces instead of all of them.
597 * systemd-networkd now implements L2TP tunnels.
599 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
600 may be used to cause autonomous and onlink prefixes received in IPv6
601 Router Advertisements to be ignored.
603 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
604 file settings may be used to tweak bridge behaviour.
606 * The new TripleSampling= option in .network files may be used to
607 configure CAN triple sampling.
609 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
610 used to point to private or preshared key for a WireGuard interface.
612 * /etc/crypttab now supports the same-cpu-crypt and
613 submit-from-crypt-cpus options to tweak encryption work scheduling
616 * systemd-tmpfiles will now take a BSD file lock before operating on a
617 contents of directory. This may be used to temporarily exclude
618 directories from aging by taking the same lock (useful for example
619 when extracting a tarball into /tmp or /var/tmp as a privileged user,
620 which might create files with really old timestamps, which
621 nevertheless should not be deleted). For further details, see:
623 https://systemd.io/TEMPORARY_DIRECTORIES
625 * systemd-tmpfiles' h line type gained support for the
626 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
627 controlling project quota inheritance.
629 * sd-boot and bootctl now implement support for an Extended Boot Loader
630 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
631 addition to the ESP partition mounted to /efi or /boot/efi.
632 Configuration file fragments, kernels, initrds and other EFI images
633 to boot will be loaded from both the ESP and XBOOTLDR partitions.
634 The XBOOTLDR partition was previously described by the Boot Loader
635 Specification, but implementation was missing in sd-boot. Support for
636 this concept allows using the sd-boot boot loader in more
637 conservative scenarios where the boot loader itself is placed in the
638 ESP but the kernels to boot (and their metadata) in a separate
641 * A system may now be booted with systemd.volatile=overlay on the
642 kernel command line, which causes the root file system to be set up
643 an overlayfs mount combining the root-only root directory with a
644 writable tmpfs. In this setup, the underlying root device is not
645 modified, and any changes are lost at reboot.
647 * Similar, systemd-nspawn can now boot containers with a volatile
648 overlayfs root with the new --volatile=overlay switch.
650 * systemd-nspawn can now consume OCI runtime bundles using a new
651 --oci-bundle= option. This implementation is fully usable, with most
652 features in the specification implemented, but since this a lot of
653 new code and functionality, this feature should most likely not
654 be used in production yet.
656 * systemd-nspawn now supports various options described by the OCI
657 runtime specification on the command-line and in .nspawn files:
658 --inaccessible=/Inaccessible= may be used to mask parts of the file
659 system tree, --console=/--pipe may be used to configure how standard
660 input, output, and error are set up.
662 * busctl learned the `emit` verb to generate D-Bus signals.
664 * systemd-analyze cat-config may be used to gather and display
665 configuration spread over multiple files, for example system and user
666 presets, tmpfiles.d, sysusers.d, udev rules, etc.
668 * systemd-analyze calendar now takes an optional new parameter
669 --iterations= which may be used to show a maximum number of iterations
670 the specified expression will elapse next.
672 * The sd-bus C API gained support for naming method parameters in the
675 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
676 the reboot() system call expects.
678 * journalctl learnt a new --cursor-file= option that points to a file
679 from which a cursor should be loaded in the beginning and to which
680 the updated cursor should be stored at the end.
682 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
683 detected by systemd-detect-virt (and may also be used in
684 ConditionVirtualization=).
686 * The behaviour of systemd-logind may now be modified with environment
687 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
688 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
689 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
690 skip the relevant operation completely (when set to false), or to
691 create a flag file in /run/systemd (when set to true), instead of
692 actually commencing the real operation when requested. The presence
693 of /run/systemd/reboot-to-firmware-setup,
694 /run/systemd/reboot-to-boot-loader-menu, and
695 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
696 boot loader implementations to replace some steps logind performs
697 during reboot with their own operations.
699 * systemctl can be used to request a reboot into the boot loader menu
700 or a specific boot loader entry with the new --boot-load-menu= and
701 --boot-loader-entry= options to a reboot command. (This requires a
702 boot loader that supports this, for example sd-boot.)
704 * kernel-install will no longer unconditionally create the output
705 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
706 snippets, but will do only if the machine-specific parent directory
707 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
708 to create this parent directory during sd-boot installation.
710 This makes it easier to use kernel-install with plugins which support
711 a different layout of the bootloader partitions (for example grub2).
713 * During package installation (with `ninja install`), we would create
714 symlinks for getty@tty1.service, systemd-networkd.service,
715 systemd-networkd.socket, systemd-resolved.service,
716 remote-cryptsetup.target, remote-fs.target,
717 systemd-networkd-wait-online.service, and systemd-timesyncd.service
718 in /etc, as if `systemctl enable` was called for those units, to make
719 the system usable immediately after installation. Now this is not
720 done anymore, and instead calling `systemctl preset-all` is
721 recommended after the first installation of systemd.
723 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
724 is built on seccomp. When turned on creation of SUID/SGID files is
727 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
728 implied if DynamicUser= is turned on for a service. This hardens
729 these services, so that they neither can benefit from nor create
730 SUID/SGID executables. This is a minor compatibility breakage, given
731 that when DynamicUser= was first introduced SUID/SGID behaviour was
732 unaffected. However, the security benefit of these two options is
733 substantial, and the setting is still relatively new, hence we opted
734 to make it mandatory for services with dynamic users.
736 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
737 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
738 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
739 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
740 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
741 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
742 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
743 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
744 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
745 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
746 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
747 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
748 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
749 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
750 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
751 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
752 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
753 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
759 * The default locale can now be configured at compile time. Otherwise,
760 a suitable default will be selected automatically (one of C.UTF-8,
763 * The version string shown by systemd and other tools now includes the
764 git commit hash when built from git. An override may be specified
765 during compilation, which is intended to be used by distributions to
766 include the package release information.
768 * systemd-cat can now filter standard input and standard error streams
769 for different syslog priorities using the new --stderr-priority=
772 * systemd-journald and systemd-journal-remote reject entries which
773 contain too many fields (CVE-2018-16865) and set limits on the
774 process' command line length (CVE-2018-16864).
776 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
779 * A new network device NamePolicy "keep" is implemented for link files,
780 and used by default in 99-default.link (the fallback configuration
781 provided by systemd). With this policy, if the network device name
782 was already set by userspace, the device will not be renamed again.
783 This matches the naming scheme that was implemented before
784 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
785 is also enabled by default, even if not specified. Effectively, this
786 means that if naming-scheme >= 240 is specified, network devices will
787 be renamed according to the configuration, even if they have been
788 renamed already, if "keep" is not specified as the naming policy in
789 the .link file. The 99-default.link file provided by systemd includes
790 "keep" for backwards compatibility, but it is recommended for user
791 installed .link files to *not* include it.
793 The "kernel" policy, which keeps kernel names declared to be
794 "persistent", now works again as documented.
796 * kernel-install script now optionally takes the paths to one or more
797 initrd files, and passes them to all plugins.
799 * The mincore() system call has been dropped from the @system-service
800 system call filter group, as it is pretty exotic and may potentially
801 used for side-channel attacks.
803 * -fPIE is dropped from compiler and linker options. Please specify
804 -Db_pie=true option to meson to build position-independent
805 executables. Note that the meson option is supported since meson-0.49.
807 * The fs.protected_regular and fs.protected_fifos sysctls, which were
808 added in Linux 4.19 to make some data spoofing attacks harder, are
809 now enabled by default. While this will hopefully improve the
810 security of most installations, it is technically a backwards
811 incompatible change; to disable these sysctls again, place the
812 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
814 fs.protected_regular = 0
815 fs.protected_fifos = 0
817 Note that the similar hardlink and symlink protection has been
818 enabled since v199, and may be disabled likewise.
820 * The files read from the EnvironmentFile= setting in unit files now
821 parse backslashes inside quotes literally, matching the behaviour of
824 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
825 now automatically become NOPs when run in a chroot() environment.
827 * The tmpfiles.d/ "C" line type will now copy directory trees not only
828 when the destination is so far missing, but also if it already exists
829 as a directory and is empty. This is useful to cater for systems
830 where directory trees are put together from multiple separate mount
831 points but otherwise empty.
833 * A new function sd_bus_close_unref() (and the associated
834 sd_bus_close_unrefp()) has been added to libsystemd, that combines
835 sd_bus_close() and sd_bus_unref() in one.
837 * udevadm control learnt a new option for --ping for testing whether a
838 systemd-udevd instance is running and reacting.
840 * udevadm trigger learnt a new option for --wait-daemon for waiting
841 systemd-udevd daemon to be initialized.
843 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
844 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
845 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
846 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
847 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
848 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
849 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
850 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
851 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
852 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
853 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
854 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
855 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
856 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
857 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
858 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
859 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
865 * NoNewPrivileges=yes has been set for all long-running services
866 implemented by systemd. Previously, this was problematic due to
867 SELinux (as this would also prohibit the transition from PID1's label
868 to the service's label). This restriction has since been lifted, but
869 an SELinux policy update is required.
870 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
872 * DynamicUser=yes is dropped from systemd-networkd.service,
873 systemd-resolved.service and systemd-timesyncd.service, which was
874 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
875 and since v236 for systemd-timesyncd.service. The users and groups
876 systemd-network, systemd-resolve and systemd-timesync are created
877 by systemd-sysusers again. Distributors or system administrators
878 may need to create these users and groups if they not exist (or need
879 to re-enable DynamicUser= for those units) while upgrading systemd.
880 Also, the clock file for systemd-timesyncd may need to move from
881 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
883 * When unit files are loaded from disk, previously systemd would
884 sometimes (depending on the unit loading order) load units from the
885 target path of symlinks in .wants/ or .requires/ directories of other
886 units. This meant that unit could be loaded from different paths
887 depending on whether the unit was requested explicitly or as a
888 dependency of another unit, not honouring the priority of directories
889 in search path. It also meant that it was possible to successfully
890 load and start units which are not found in the unit search path, as
891 long as they were requested as a dependency and linked to from
892 .wants/ or .requires/. The target paths of those symlinks are not
893 used for loading units anymore and the unit file must be found in
896 * A new service type has been added: Type=exec. It's very similar to
897 Type=simple but ensures the service manager will wait for both fork()
898 and execve() of the main service binary to complete before proceeding
899 with follow-up units. This is primarily useful so that the manager
900 propagates any errors in the preparation phase of service execution
901 back to the job that requested the unit to be started. For example,
902 consider a service that has ExecStart= set to a file system binary
903 that doesn't exist. With Type=simple starting the unit would be
904 considered instantly successful, as only fork() has to complete
905 successfully and the manager does not wait for execve(), and hence
906 its failure is seen "too late". With the new Type=exec service type
907 starting the unit will fail, as the manager will wait for the
908 execve() and notice its failure, which is then propagated back to the
911 NOTE: with the next release 241 of systemd we intend to change the
912 systemd-run tool to default to Type=exec for transient services
913 started by it. This should be mostly safe, but in specific corner
914 cases might result in problems, as the systemd-run tool will then
915 block on NSS calls (such as user name look-ups due to User=) done
916 between the fork() and execve(), which under specific circumstances
917 might cause problems. It is recommended to specify "-p Type=simple"
918 explicitly in the few cases where this applies. For regular,
919 non-transient services (i.e. those defined with unit files on disk)
920 we will continue to default to Type=simple.
922 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
923 userspace processes is set to 1024 (soft) and 4096
924 (hard). Previously, systemd passed this on unmodified to all
925 processes it forked off. With this systemd release the hard limit
926 systemd passes on is increased to 512K, overriding the kernel's
927 defaults and substantially increasing the number of simultaneous file
928 descriptors unprivileged userspace processes can allocate. Note that
929 the soft limit remains at 1024 for compatibility reasons: the
930 traditional UNIX select() call cannot deal with file descriptors >=
931 1024 and increasing the soft limit globally might thus result in
932 programs unexpectedly allocating a high file descriptor and thus
933 failing abnormally when attempting to use it with select() (of
934 course, programs shouldn't use select() anymore, and prefer
935 poll()/epoll, but the call unfortunately remains undeservedly popular
936 at this time). This change reflects the fact that file descriptor
937 handling in the Linux kernel has been optimized in more recent
938 kernels and allocating large numbers of them should be much cheaper
939 both in memory and in performance than it used to be. Programs that
940 want to take benefit of the increased limit have to "opt-in" into
941 high file descriptors explicitly by raising their soft limit. Of
942 course, when they do that they must acknowledge that they cannot use
943 select() anymore (and neither can any shared library they use — or
944 any shared library used by any shared library they use and so on).
945 Which default hard limit is most appropriate is of course hard to
946 decide. However, given reports that ~300K file descriptors are used
947 in real-life applications we believe 512K is sufficiently high as new
948 default for now. Note that there are also reports that using very
949 high hard limits (e.g. 1G) is problematic: some software allocates
950 large arrays with one element for each potential file descriptor
951 (Java, …) — a high hard limit thus triggers excessively large memory
952 allocations in these applications. Hopefully, the new default of 512K
953 is a good middle ground: higher than what real-life applications
954 currently need, and low enough for avoid triggering excessively large
955 allocations in problematic software. (And yes, somebody should fix
958 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
959 to the highest possible values, as separate accounting of file
960 descriptors is no longer necessary, as memcg tracks them correctly as
961 part of the memory accounting anyway. Thus, from the four limits on
962 file descriptors currently enforced (fs.file-max, fs.nr_open,
963 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
964 and keep only the latter two. A set of build-time options
965 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
966 has been added to revert this change in behaviour, which might be
967 an option for systems that turn off memcg in the kernel.
969 * When no /etc/locale.conf file exists (and hence no locale settings
970 are in place), systemd will now use the "C.UTF-8" locale by default,
971 and set LANG= to it. This locale is supported by various
972 distributions including Fedora, with clear indications that upstream
973 glibc is going to make it available too. This locale enables UTF-8
974 mode by default, which appears appropriate for 2018.
976 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
977 default. This effectively switches the RFC3704 Reverse Path filtering
978 from Strict mode to Loose mode. This is more appropriate for hosts
979 that have multiple links with routes to the same networks (e.g.
980 a client with a Wi-Fi and Ethernet both connected to the internet).
982 Consult the kernel documentation for details on this sysctl:
983 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
985 * CPUAccounting=yes no longer enables the CPU controller when using
986 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
987 statistics are now provided independently from the CPU controller.
989 * Support for disabling a particular cgroup controller within a sub-tree
990 has been added through the DisableControllers= directive.
992 * cgroup_no_v1=all on the kernel command line now also implies
993 using the unified cgroup hierarchy, unless one explicitly passes
994 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
996 * The new "MemoryMin=" unit file property may now be used to set the
997 memory usage protection limit of processes invoked by the unit. This
998 controls the cgroup v2 memory.min attribute. Similarly, the new
999 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
1000 cgroup v2 io.latency cgroup property for configuring per-service I/O
1003 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
1004 to the cgroup v1 "devices" cgroup controller.
1006 * systemd-escape now is able to combine --unescape with --template. It
1007 also learnt a new option --instance for extracting and unescaping the
1008 instance part of a unit name.
1010 * sd-bus now provides the sd_bus_message_readv() which is similar to
1011 sd_bus_message_read() but takes a va_list object. The pair
1012 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
1013 has been added for configuring the default method call timeout to
1014 use. sd_bus_error_move() may be used to efficiently move the contents
1015 from one sd_bus_error structure to another, invalidating the
1016 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
1017 be used to control whether a bus connection object is automatically
1018 flushed when an sd-event loop is exited.
1020 * When processing classic BSD syslog log messages, journald will now
1021 save the original time-stamp string supplied in the new
1022 SYSLOG_TIMESTAMP= journal field. This permits consumers to
1023 reconstruct the original BSD syslog message more correctly.
1025 * StandardOutput=/StandardError= in service files gained support for
1026 new "append:…" parameters, for connecting STDOUT/STDERR of a service
1027 to a file, and appending to it.
1029 * The signal to use as last step of killing of unit processes is now
1030 configurable. Previously it was hard-coded to SIGKILL, which may now
1031 be overridden with the new KillSignal= setting. Note that this is the
1032 signal used when regular termination (i.e. SIGTERM) does not suffice.
1033 Similarly, the signal used when aborting a program in case of a
1034 watchdog timeout may now be configured too (WatchdogSignal=).
1036 * The XDG_SESSION_DESKTOP environment variable may now be configured in
1037 the pam_systemd argument line, using the new desktop= switch. This is
1038 useful to initialize it properly from a display manager without
1039 having to touch C code.
1041 * Most configuration options that previously accepted percentage values
1042 now also accept permille values with the '‰' suffix (instead of '%').
1044 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
1047 * systemd-resolved's configuration file resolved.conf gained a new
1048 option ReadEtcHosts= which may be used to turn off processing and
1049 honoring /etc/hosts entries.
1051 * The "--wait" switch may now be passed to "systemctl
1052 is-system-running", in which case the tool will synchronously wait
1053 until the system finished start-up.
1055 * hostnamed gained a new bus call to determine the DMI product UUID.
1057 * On x86-64 systemd will now prefer using the RDRAND processor
1058 instruction over /dev/urandom whenever it requires randomness that
1059 neither has to be crypto-grade nor should be reproducible. This
1060 should substantially reduce the amount of entropy systemd requests
1061 from the kernel during initialization on such systems, though not
1062 reduce it to zero. (Why not zero? systemd still needs to allocate
1063 UUIDs and such uniquely, which require high-quality randomness.)
1065 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
1066 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
1067 for forcing the "Other Information" bit in IPv6 RA messages. The
1068 bonding logic gained four new options AdActorSystemPriority=,
1069 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
1070 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
1071 shuffling of flows. The tunnel logic gained a new
1072 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
1073 Deployment. The policy rule logic gained four new options IPProtocol=,
1074 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
1075 support for the MulticastToUnicast= option. networkd also gained
1076 support for configuring static IPv4 ARP or IPv6 neighbor entries.
1078 * .preset files (as read by 'systemctl preset') may now be used to
1079 instantiate services.
1081 * /etc/crypttab now understands the sector-size= option to configure
1082 the sector size for an encrypted partition.
1084 * Key material for encrypted disks may now be placed on a formatted
1085 medium, and referenced from /etc/crypttab by the UUID of the file
1086 system, followed by "=" suffixed by the path to the key file.
1088 * The "collect" udev component has been removed without replacement, as
1089 it is neither used nor maintained.
1091 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
1092 LogsDirectory=, ConfigurationDirectory= settings are used in a
1093 service the executed processes will now receive a set of environment
1094 variables containing the full paths of these directories.
1095 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
1096 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
1097 are used. Note that these options may be used multiple times per
1098 service in which case the resulting paths will be concatenated and
1099 separated by colons.
1101 * Predictable interface naming has been extended to cover InfiniBand
1102 NICs. They will be exposed with an "ib" prefix.
1104 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
1105 which case the respective line failing is ignored.
1107 * .link files may now be used to configure the equivalent to the
1108 "ethtool advertise" commands.
1110 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
1111 alternative to libudev.h. Previously, the latter was just an internal
1112 wrapper around the former, but now these two APIs are exposed
1115 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
1116 which calculates an app-specific boot ID similar to how
1117 sd_id128_get_machine_app_specific() generates an app-specific machine
1120 * A new tool systemd-id128 has been added that can be used to determine
1121 and generate various 128bit IDs.
1123 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
1126 * systemd-hibernate-resume-generator will now honor the "noresume"
1127 kernel command line option, in which case it will bypass resuming
1128 from any hibernated image.
1130 * The systemd-sleep.conf configuration file gained new options
1131 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
1132 AllowHybridSleep= for prohibiting specific sleep modes even if the
1133 kernel exports them.
1135 * portablectl is now officially supported and has thus moved to
1138 * bootctl learnt the two new commands "set-default" and "set-oneshot"
1139 for setting the default boot loader item to boot to (either
1140 persistently or only for the next boot). This is currently only
1141 compatible with sd-boot, but may be implemented on other boot loaders
1142 too, that follow the boot loader interface. The updated interface is
1143 now documented here:
1145 https://systemd.io/BOOT_LOADER_INTERFACE
1147 * A new kernel command line option systemd.early_core_pattern= is now
1148 understood which may be used to influence the core_pattern PID 1
1149 installs during early boot.
1151 * busctl learnt two new options -j and --json= for outputting method
1152 call replies, properties and monitoring output in JSON.
1154 * journalctl's JSON output now supports simple ANSI coloring as well as
1155 a new "json-seq" mode for generating RFC7464 output.
1157 * Unit files now support the %g/%G specifiers that resolve to the UNIX
1158 group/GID of the service manager runs as, similar to the existing
1159 %u/%U specifiers that resolve to the UNIX user/UID.
1161 * systemd-logind learnt a new global configuration option
1162 UserStopDelaySec= that may be set in logind.conf. It specifies how
1163 long the systemd --user instance shall remain started after a user
1164 logs out. This is useful to speed up repetitive re-connections of the
1165 same user, as it means the user's service manager doesn't have to be
1166 stopped/restarted on each iteration, but can be reused between
1167 subsequent options. This setting defaults to 10s. systemd-logind also
1168 exports two new properties on its Manager D-Bus objects indicating
1169 whether the system's lid is currently closed, and whether the system
1172 * systemd gained support for a generic boot counting logic, which
1173 generically permits automatic reverting to older boot loader entries
1174 if newer updated ones don't work. The boot loader side is implemented
1175 in sd-boot, but is kept open for other boot loaders too. For details
1178 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
1180 * The SuccessAction=/FailureAction= unit file settings now learnt two
1181 new parameters: "exit" and "exit-force", which result in immediate
1182 exiting of the service manager, and are only useful in systemd --user
1183 and container environments.
1185 * Unit files gained support for a pair of options
1186 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
1187 exit status to use as service manager exit status when
1188 SuccessAction=/FailureAction= is set to exit or exit-force.
1190 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
1191 options may now be used to configure the log rate limiting applied by
1192 journald per-service.
1194 * systemd-analyze gained a new verb "timespan" for parsing and
1195 normalizing time span values (i.e. strings like "5min 7s 8us").
1197 * systemd-analyze also gained a new verb "security" for analyzing the
1198 security and sand-boxing settings of services in order to determine an
1199 "exposure level" for them, indicating whether a service would benefit
1200 from more sand-boxing options turned on for them.
1202 * "systemd-analyze syscall-filter" will now also show system calls
1203 supported by the local kernel but not included in any of the defined
1206 * .nspawn files now understand the Ephemeral= setting, matching the
1207 --ephemeral command line switch.
1209 * sd-event gained the new APIs sd_event_source_get_floating() and
1210 sd_event_source_set_floating() for controlling whether a specific
1211 event source is "floating", i.e. destroyed along with the even loop
1214 * Unit objects on D-Bus gained a new "Refs" property that lists all
1215 clients that currently have a reference on the unit (to ensure it is
1218 * The JoinControllers= option in system.conf is no longer supported, as
1219 it didn't work correctly, is hard to support properly, is legacy (as
1220 the concept only exists on cgroup v1) and apparently wasn't used.
1222 * Journal messages that are generated whenever a unit enters the failed
1223 state are now tagged with a unique MESSAGE_ID. Similarly, messages
1224 generated whenever a service process exits are now made recognizable,
1225 too. A tagged message is also emitted whenever a unit enters the
1226 "dead" state on success.
1228 * systemd-run gained a new switch --working-directory= for configuring
1229 the working directory of the service to start. A shortcut -d is
1230 equivalent, setting the working directory of the service to the
1231 current working directory of the invoking program. The new --shell
1232 (or just -S) option has been added for invoking the $SHELL of the
1233 caller as a service, and implies --pty --same-dir --wait --collect
1234 --service-type=exec. Or in other words, "systemd-run -S" is now the
1235 quickest way to quickly get an interactive in a fully clean and
1236 well-defined system service context.
1238 * machinectl gained a new verb "import-fs" for importing an OS tree
1239 from a directory. Moreover, when a directory or tarball is imported
1240 and single top-level directory found with the OS itself below the OS
1241 tree is automatically mangled and moved one level up.
1243 * systemd-importd will no longer set up an implicit btrfs loop-back
1244 file system on /var/lib/machines. If one is already set up, it will
1245 continue to be used.
1247 * A new generator "systemd-run-generator" has been added. It will
1248 synthesize a unit from one or more program command lines included in
1249 the kernel command line. This is very useful in container managers
1252 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
1254 This will run "systemd-nspawn" on an image, invoke the specified
1255 command line and immediately shut down the container again, returning
1256 the command line's exit code.
1258 * The block device locking logic is now documented:
1260 https://systemd.io/BLOCK_DEVICE_LOCKING
1262 * loginctl and machinectl now optionally output the various tables in
1263 JSON using the --output= switch. It is our intention to add similar
1264 support to systemctl and all other commands.
1266 * udevadm's query and trigger verb now optionally take a .device unit
1269 * systemd-udevd's network naming logic now understands a new
1270 net.naming-scheme= kernel command line switch, which may be used to
1271 pick a specific version of the naming scheme. This helps stabilizing
1272 interface names even as systemd/udev are updated and the naming logic
1275 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
1276 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
1277 initialize one to all 0xFF.
1279 * After loading the SELinux policy systemd will now recursively relabel
1280 all files and directories listed in
1281 /run/systemd/relabel-extra.d/*.relabel (which should be simple
1282 newline separated lists of paths) in addition to the ones it already
1283 implicitly relabels in /run, /dev and /sys. After the relabelling is
1284 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
1285 removed. This is useful to permit initrds (i.e. code running before
1286 the SELinux policy is in effect) to generate files in the host
1287 filesystem safely and ensure that the correct label is applied during
1288 the transition to the host OS.
1290 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
1291 mknod() handling in user namespaces. Previously mknod() would always
1292 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
1293 but device nodes generated that way cannot be opened, and attempts to
1294 open them result in EPERM. This breaks the "graceful fallback" logic
1295 in systemd's PrivateDevices= sand-boxing option. This option is
1296 implemented defensively, so that when systemd detects it runs in a
1297 restricted environment (such as a user namespace, or an environment
1298 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
1299 where device nodes cannot be created the effect of PrivateDevices= is
1300 bypassed (following the logic that 2nd-level sand-boxing is not
1301 essential if the system systemd runs in is itself already sand-boxed
1302 as a whole). This logic breaks with 4.18 in container managers where
1303 user namespacing is used: suddenly PrivateDevices= succeeds setting
1304 up a private /dev/ file system containing devices nodes — but when
1305 these are opened they don't work.
1307 At this point it is recommended that container managers utilizing
1308 user namespaces that intend to run systemd in the payload explicitly
1309 block mknod() with seccomp or similar, so that the graceful fallback
1312 We are very sorry for the breakage and the requirement to change
1313 container configurations for newer kernels. It's purely caused by an
1314 incompatible kernel change. The relevant kernel developers have been
1315 notified about this userspace breakage quickly, but they chose to
1318 * PermissionsStartOnly= setting is deprecated (but is still supported
1319 for backwards compatibility). The same functionality is provided by
1320 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
1323 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
1324 pam_systemd anymore.
1326 * The naming scheme for network devices was changed to always rename
1327 devices, even if they were already renamed by userspace. The "kernel"
1328 policy was changed to only apply as a fallback, if no other naming
1331 * The requirements to build systemd is bumped to meson-0.46 and
1334 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
1335 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
1336 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
1337 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
1338 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
1339 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
1340 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
1341 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
1342 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
1343 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
1344 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
1345 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
1346 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
1347 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
1348 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
1349 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
1350 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
1351 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
1352 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
1353 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
1354 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
1355 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
1356 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
1357 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
1358 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
1359 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
1360 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
1361 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
1362 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
1363 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
1364 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
1365 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
1366 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
1367 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
1368 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
1369 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
1370 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
1371 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
1372 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
1373 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
1374 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
1375 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
1376 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
1377 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
1378 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
1380 — Warsaw, 2018-12-21
1384 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
1385 builtin will name network interfaces differently than in previous
1386 versions for virtual network interfaces created with SR-IOV and NPAR
1387 and for devices where the PCI network controller device does not have
1388 a slot number associated.
1390 SR-IOV virtual devices are now named based on the name of the parent
1391 interface, with a suffix of "v<N>", where <N> is the virtual device
1392 number. Previously those virtual devices were named as if completely
1395 The ninth and later NPAR virtual devices will be named following the
1396 scheme used for the first eight NPAR partitions. Previously those
1397 devices were not renamed and the kernel default (eth<n>) was used.
1399 "net_id" will also generate names for PCI devices where the PCI
1400 network controller device does not have an associated slot number
1401 itself, but one of its parents does. Previously those devices were
1402 not renamed and the kernel default (eth<n>) was used.
1404 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
1405 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
1406 the unit. So, it is expected that the default behavior of
1407 systemd-logind is not changed. However, if distribution packagers or
1408 administrators disabled or modified IPAddressDeny= setting by a
1409 drop-in config file, then it may be necessary to update the file to
1410 re-enable AF_INET and AF_INET6 to support network user name services,
1413 * When the RestrictNamespaces= unit property is specified multiple
1414 times, then the specified types are merged now. Previously, only the
1415 last assignment was used. So, if distribution packagers or
1416 administrators modified the setting by a drop-in config file, then it
1417 may be necessary to update the file.
1419 * When OnFailure= is used in combination with Restart= on a service
1420 unit, then the specified units will no longer be triggered on
1421 failures that result in restarting. Previously, the specified units
1422 would be activated each time the unit failed, even when the unit was
1423 going to be restarted automatically. This behaviour contradicted the
1424 documentation. With this release the code is adjusted to match the
1427 * systemd-tmpfiles will now print a notice whenever it encounters
1428 tmpfiles.d/ lines referencing the /var/run/ directory. It will
1429 recommend reworking them to use the /run/ directory instead (for
1430 which /var/run/ is simply a symlinked compatibility alias). This way
1431 systemd-tmpfiles can properly detect line conflicts and merge lines
1432 referencing the same file by two paths, without having to access
1435 * systemctl disable/unmask/preset/preset-all cannot be used with
1436 --runtime. Previously this was allowed, but resulted in unintuitive
1437 behaviour that wasn't useful. systemctl disable/unmask will now undo
1438 both runtime and persistent enablement/masking, i.e. it will remove
1439 any relevant symlinks both in /run and /etc.
1441 * Note that all long-running system services shipped with systemd will
1442 now default to a system call whitelist (rather than a blacklist, as
1443 before). In particular, systemd-udevd will now enforce one too. For
1444 most cases this should be safe, however downstream distributions
1445 which disabled sandboxing of systemd-udevd (specifically the
1446 MountFlags= setting), might want to disable this security feature
1447 too, as the default whitelisting will prohibit all mount, swap,
1448 reboot and clock changing operations from udev rules.
1450 * sd-boot acquired new loader configuration settings to optionally turn
1451 off Windows and MacOS boot partition discovery as well as
1452 reboot-into-firmware menu items. It is also able to pick a better
1453 screen resolution for HiDPI systems, and now provides loader
1454 configuration settings to change the resolution explicitly.
1456 * systemd-resolved now supports DNS-over-TLS. It's still
1457 turned off by default, use DNSOverTLS=opportunistic to turn it on in
1458 resolved.conf. We intend to make this the default as soon as couple
1459 of additional techniques for optimizing the initial latency caused by
1460 establishing a TLS/TCP connection are implemented.
1462 * systemd-resolved.service and systemd-networkd.service now set
1463 DynamicUser=yes. The users systemd-resolve and systemd-network are
1464 not created by systemd-sysusers anymore.
1466 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
1467 that embed a network facing module into any process using getpwuid()
1468 or related call: the dynamic allocation of the user ID for
1469 systemd-resolved.service means the service manager has to check NSS
1470 if the user name is already taken when forking off the service. Since
1471 the user in the common case won't be defined in /etc/passwd the
1472 lookup is likely to trigger nss-ldap which in turn might use NSS to
1473 ask systemd-resolved for hostname lookups. This will hence result in
1474 a deadlock: a user name lookup in order to start
1475 systemd-resolved.service will result in a host name lookup for which
1476 systemd-resolved.service needs to be started already. There are
1477 multiple ways to work around this problem: pre-allocate the
1478 "systemd-resolve" user on such systems, so that nss-ldap won't be
1479 triggered; or use a different NSS package that doesn't do networking
1480 in-process but provides a local asynchronous name cache; or configure
1481 the NSS package to avoid lookups for UIDs in the range `pkg-config
1482 systemd --variable=dynamicuidmin` … `pkg-config systemd
1483 --variable=dynamicuidmax`, so that it does not consider itself
1484 authoritative for the same UID range systemd allocates dynamic users
1487 * The systemd-resolve tool has been renamed to resolvectl (it also
1488 remains available under the old name, for compatibility), and its
1489 interface is now verb-based, similar in style to the other <xyz>ctl
1490 tools, such as systemctl or loginctl.
1492 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
1493 compatibility. It may be symlinked under the 'resolvconf' name, in
1494 which case it will take arguments and input compatible with the
1495 Debian and FreeBSD resolvconf tool.
1497 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
1498 where the system initially suspends, and after a timeout resumes and
1501 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
1502 set the client will only send a DUID as client identifier.
1504 * The nss-systemd glibc NSS module will now enumerate dynamic users and
1505 groups in effect. Previously, it could resolve UIDs/GIDs to user
1506 names/groups and vice versa, but did not support enumeration.
1508 * journald's Compress= configuration setting now optionally accepts a
1509 byte threshold value. All journal objects larger than this threshold
1510 will be compressed, smaller ones will not. Previously this threshold
1511 was not configurable and set to 512.
1513 * A new system.conf setting NoNewPrivileges= is now available which may
1514 be used to turn off acquisition of new privileges system-wide
1515 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
1516 for all its children). Note that turning this option on means setuid
1517 binaries and file system capabilities lose their special powers.
1518 While turning on this option is a big step towards a more secure
1519 system, doing so is likely to break numerous pre-existing UNIX tools,
1520 in particular su and sudo.
1522 * A new service systemd-time-sync-wait.service has been added. If
1523 enabled it will delay the time-sync.target unit at boot until time
1524 synchronization has been received from the network. This
1525 functionality is useful on systems lacking a local RTC or where it is
1526 acceptable that the boot process shall be delayed by external network
1529 * When hibernating, systemd will now inform the kernel of the image
1530 write offset, on kernels new enough to support this. This means swap
1531 files should work for hibernation now.
1533 * When loading unit files, systemd will now look for drop-in unit files
1534 extensions in additional places. Previously, for a unit file name
1535 "foo-bar-baz.service" it would look for dropin files in
1536 "foo-bar-baz.service.d/*.conf". Now, it will also look in
1537 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
1538 service name truncated after all inner dashes. This scheme allows
1539 writing drop-ins easily that apply to a whole set of unit files at
1540 once. It's particularly useful for mount and slice units (as their
1541 naming is prefix based), but is also useful for service and other
1542 units, for packages that install multiple unit files at once,
1543 following a strict naming regime of beginning the unit file name with
1544 the package's name. Two new specifiers are now supported in unit
1545 files to match this: %j and %J are replaced by the part of the unit
1546 name following the last dash.
1548 * Unit files and other configuration files that support specifier
1549 expansion now understand another three new specifiers: %T and %V will
1550 resolve to /tmp and /var/tmp respectively, or whatever temporary
1551 directory has been set for the calling user. %E will expand to either
1552 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
1554 * The ExecStart= lines of unit files are no longer required to
1555 reference absolute paths. If non-absolute paths are specified the
1556 specified binary name is searched within the service manager's
1557 built-in $PATH, which may be queried with 'systemd-path
1558 search-binaries-default'. It's generally recommended to continue to
1559 use absolute paths for all binaries specified in unit files.
1561 * Units gained a new load state "bad-setting", which is used when a
1562 unit file was loaded, but contained fatal errors which prevent it
1563 from being started (for example, a service unit has been defined
1564 lacking both ExecStart= and ExecStop= lines).
1566 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
1567 support alternative debuggers, for example lldb. The old name
1568 continues to be available however, for compatibility reasons. Use the
1569 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
1570 to pick an alternative debugger instead of the default gdb.
1572 * systemctl and the other tools will now output escape sequences that
1573 generate proper clickable hyperlinks in various terminal emulators
1574 where useful (for example, in the "systemctl status" output you can
1575 now click on the unit file name to quickly open it in the
1576 editor/viewer of your choice). Note that not all terminal emulators
1577 support this functionality yet, but many do. Unfortunately, the
1578 "less" pager doesn't support this yet, hence this functionality is
1579 currently automatically turned off when a pager is started (which
1580 happens quite often due to auto-paging). We hope to remove this
1581 limitation as soon as "less" learns these escape sequences. This new
1582 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
1583 environment variable. For details on these escape sequences see:
1584 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
1586 * networkd's .network files now support a new IPv6MTUBytes= option for
1587 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
1588 option in the [Route] section to configure the MTU to use for
1589 specific routes. It also gained support for configuration of the DHCP
1590 "UserClass" option through the new UserClass= setting. It gained
1591 three new options in the new [CAN] section for configuring CAN
1592 networks. The MULTICAST and ALLMULTI interface flags may now be
1593 controlled explicitly with the new Multicast= and AllMulticast=
1596 * networkd will now automatically make use of the kernel's route
1597 expiration feature, if it is available.
1599 * udevd's .link files now support setting the number of receive and
1600 transmit channels, using the RxChannels=, TxChannels=,
1601 OtherChannels=, CombinedChannels= settings.
1603 * Support for UDPSegmentationOffload= has been removed, given its
1604 limited support in hardware, and waning software support.
1606 * networkd's .netdev files now support creating "netdevsim" interfaces.
1608 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
1609 to query the unit belonging to a specific kernel control group.
1611 * systemd-analyze gained a new verb "cat-config", which may be used to
1612 dump the contents of any configuration file, with all its matching
1613 drop-in files added in, and honouring the usual search and masking
1614 logic applied to systemd configuration files. For example use
1615 "systemd-analyze cat-config systemd/system.conf" to get the complete
1616 system configuration file of systemd how it would be loaded by PID 1
1617 itself. Similar to this, various tools such as systemd-tmpfiles or
1618 systemd-sysusers, gained a new option "--cat-config", which does the
1619 corresponding operation for their own configuration settings. For
1620 example, "systemd-tmpfiles --cat-config" will now output the full
1621 list of tmpfiles.d/ lines in place.
1623 * timedatectl gained three new verbs: "show" shows bus properties of
1624 systemd-timedated, "timesync-status" shows the current NTP
1625 synchronization state of systemd-timesyncd, and "show-timesync"
1626 shows bus properties of systemd-timesyncd.
1628 * systemd-timesyncd gained a bus interface on which it exposes details
1631 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
1632 understood by systemd-timedated. It takes a colon-separated list of
1633 unit names of NTP client services. The list is used by
1634 "timedatectl set-ntp".
1636 * systemd-nspawn gained a new --rlimit= switch for setting initial
1637 resource limits for the container payload. There's a new switch
1638 --hostname= to explicitly override the container's hostname. A new
1639 --no-new-privileges= switch may be used to control the
1640 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
1641 --oom-score-adjust= switch controls the OOM scoring adjustment value
1642 for the payload. The new --cpu-affinity= switch controls the CPU
1643 affinity of the container payload. The new --resolv-conf= switch
1644 allows more detailed control of /etc/resolv.conf handling of the
1645 container. Similarly, the new --timezone= switch allows more detailed
1646 control of /etc/localtime handling of the container.
1648 * systemd-detect-virt gained a new --list switch, which will print a
1649 list of all currently known VM and container environments.
1651 * Support for "Portable Services" has been added, see
1652 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
1653 experimental, but this is expected to change soon. Reflecting this
1654 experimental state, the "portablectl" binary is not installed into
1655 /usr/bin yet. The binary has to be called with the full path
1656 /usr/lib/systemd/portablectl instead.
1658 * journalctl's and systemctl's -o switch now knows a new log output
1659 mode "with-unit". The output it generates is very similar to the
1660 regular "short" mode, but displays the unit name instead of the
1661 syslog tag for each log line. Also, the date is shown with timezone
1662 information. This mode is probably more useful than the classic
1663 "short" output mode for most purposes, except where pixel-perfect
1664 compatibility with classic /var/log/messages formatting is required.
1666 * A new --dump-bus-properties switch has been added to the systemd
1667 binary, which may be used to dump all supported D-Bus properties.
1668 (Options which are still supported, but are deprecated, are *not*
1671 * sd-bus gained a set of new calls:
1672 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
1673 enable/disable the "floating" state of a bus slot object,
1674 i.e. whether the slot object pins the bus it is allocated for into
1675 memory or if the bus slot object gets disconnected when the bus goes
1676 away. sd_bus_open_with_description(),
1677 sd_bus_open_user_with_description(),
1678 sd_bus_open_system_with_description() may be used to allocate bus
1679 objects and set their description string already during allocation.
1681 * sd-event gained support for watching inotify events from the event
1682 loop, in an efficient way, sharing inotify handles between multiple
1683 users. For this a new function sd_event_add_inotify() has been added.
1685 * sd-event and sd-bus gained support for calling special user-supplied
1686 destructor functions for userdata pointers associated with
1687 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
1688 functions sd_bus_slot_set_destroy_callback,
1689 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
1690 sd_bus_track_get_destroy_callback,
1691 sd_event_source_set_destroy_callback,
1692 sd_event_source_get_destroy_callback have been added.
1694 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
1696 * PID 1 will now automatically reschedule .timer units whenever the
1697 local timezone changes. (They previously got rescheduled
1698 automatically when the system clock changed.)
1700 * New documentation has been added to document cgroups delegation,
1701 portable services and the various code quality tools we have set up:
1703 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
1704 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
1705 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
1707 * The Boot Loader Specification has been added to the source tree.
1709 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
1711 While moving it into our source tree we have updated it and further
1712 changes are now accepted through the usual github PR workflow.
1714 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
1715 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
1716 earlier PAM modules. The data in these fields is used to initialize
1717 the session scope's resource properties. Thus external PAM modules
1718 may now configure per-session limits, for example sourced from
1719 external user databases.
1721 * socket units with Accept=yes will now maintain a "refused" counter in
1722 addition to the existing "accepted" counter, counting connections
1723 refused due to the enforced limits.
1725 * The "systemd-path search-binaries-default" command may now be use to
1726 query the default, built-in $PATH PID 1 will pass to the services it
1729 * A new unit file setting PrivateMounts= has been added. It's a boolean
1730 option. If enabled the unit's processes are invoked in their own file
1731 system namespace. Note that this behaviour is also implied if any
1732 other file system namespacing options (such as PrivateTmp=,
1733 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
1734 primarily useful for services that do not use any of the other file
1735 system namespacing options. One such service is systemd-udevd.service
1736 where this is now used by default.
1738 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
1739 when the system is booted in UEFI "secure mode".
1741 * A new unit "system-update-pre.target" is added, which defines an
1742 optional synchronization point for offline system updates, as
1743 implemented by the pre-existing "system-update.target" unit. It
1744 allows ordering services before the service that executes the actual
1745 update process in a generic way.
1747 * Systemd now emits warnings whenever .include syntax is used.
1749 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
1750 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
1751 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
1752 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
1753 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
1754 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
1755 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
1756 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
1757 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
1758 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
1759 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
1760 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
1761 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
1762 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
1763 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
1764 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
1765 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
1766 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
1767 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
1768 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
1769 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
1770 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
1771 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
1772 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
1773 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
1774 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
1775 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
1776 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
1777 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1779 — Berlin, 2018-06-22
1783 * The MemoryAccounting= unit property now defaults to on. After
1784 discussions with the upstream control group maintainers we learnt
1785 that the negative impact of cgroup memory accounting on current
1786 kernels is finally relatively minimal, so that it should be safe to
1787 enable this by default without affecting system performance. Besides
1788 memory accounting only task accounting is turned on by default, all
1789 other forms of resource accounting (CPU, IO, IP) remain off for now,
1790 because it's not clear yet that their impact is small enough to move
1791 from opt-in to opt-out. We recommend downstreams to leave memory
1792 accounting on by default if kernel 4.14 or higher is primarily
1793 used. On very resource constrained systems or when support for old
1794 kernels is a necessity, -Dmemory-accounting-default=false can be used
1795 to revert this change.
1797 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
1798 %udev_rules_update) and the journal catalog (%journal_catalog_update)
1799 from the upgrade scriptlets of individual packages now do nothing.
1800 Transfiletriggers have been added which will perform those updates
1801 once at the end of the transaction.
1803 Similar transfiletriggers have been added to execute any sysctl.d
1804 and binfmt.d rules. Thus, it should be unnecessary to provide any
1805 scriptlets to execute this configuration from package installation
1808 * systemd-sysusers gained a mode where the configuration to execute is
1809 specified on the command line, but this configuration is not executed
1810 directly, but instead it is merged with the configuration on disk,
1811 and the result is executed. This is useful for package installation
1812 scripts which want to create the user before installing any files on
1813 disk (in case some of those files are owned by that user), while
1814 still allowing local admin overrides.
1816 This functionality is exposed to rpm scriptlets through a new
1817 %sysusers_create_package macro. Old %sysusers_create and
1818 %sysusers_create_inline macros are deprecated.
1820 A transfiletrigger for sysusers.d configuration is now installed,
1821 which means that it should be unnecessary to call systemd-sysusers from
1822 package installation scripts, unless the package installs any files
1823 owned by those newly-created users, in which case
1824 %sysusers_create_package should be used.
1826 * Analogous change has been done for systemd-tmpfiles: it gained a mode
1827 where the command-line configuration is merged with the configuration
1828 on disk. This is exposed as the new %tmpfiles_create_package macro,
1829 and %tmpfiles_create is deprecated. A transfiletrigger is installed
1830 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
1831 from package installation scripts.
1833 * sysusers.d configuration for a user may now also specify the group
1834 number, in addition to the user number ("u username 123:456"), or
1835 without the user number ("u username -:456").
1837 * Configution items for systemd-sysusers can now be specified as
1838 positional arguments when the new --inline switch is used.
1840 * The login shell of users created through sysusers.d may now be
1841 specified (previously, it was always /bin/sh for root and
1842 /sbin/nologin for other users).
1844 * systemd-analyze gained a new --global switch to look at global user
1845 configuration. It also gained a unit-paths verb to list the unit load
1846 paths that are compiled into systemd (which can be used with
1847 --systemd, --user, or --global).
1849 * udevadm trigger gained a new --settle/-w option to wait for any
1850 triggered events to finish (but just those, and not any other events
1851 which are triggered meanwhile).
1853 * The action that systemd-logind takes when the lid is closed and the
1854 machine is connected to external power can now be configured using
1855 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
1856 was determined by HandleLidSwitch=, and, for backwards compatibility,
1857 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
1859 * journalctl will periodically call sd_journal_process() to make it
1860 resilient against inotify queue overruns when journal files are
1861 rotated very quickly.
1863 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
1864 sd_bus_get_n_queued_write — may be used to check the number of
1865 pending bus messages.
1867 * systemd gained a new
1868 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
1869 which can be used to migrate foreign processes to scope and service
1870 units. The primary user for this new API is systemd itself: the
1871 systemd --user instance uses this call of the systemd --system
1872 instance to migrate processes if it itself gets the request to
1873 migrate processes and the kernel refuses this due to access
1874 restrictions. Thanks to this "systemd-run --scope --user …" works
1875 again in pure cgroup v2 environments when invoked from the user
1878 * A new TemporaryFileSystem= setting can be used to mask out part of
1879 the real file system tree with tmpfs mounts. This may be combined
1880 with BindPaths= and BindReadOnlyPaths= to hide files or directories
1881 not relevant to the unit, while still allowing some paths lower in
1882 the tree to be accessed.
1884 ProtectHome=tmpfs may now be used to hide user home and runtime
1885 directories from units, in a way that is mostly equivalent to
1886 "TemporaryFileSystem=/home /run/user /root".
1888 * Non-service units are now started with KeyringMode=shared by default.
1889 This means that mount and swapon and other mount tools have access
1890 to keys in the main keyring.
1892 * /sys/fs/bpf is now mounted automatically.
1894 * QNX virtualization is now detected by systemd-detect-virt and may
1895 be used in ConditionVirtualization=.
1897 * IPAccounting= may now be enabled also for slice units.
1899 * A new -Dsplit-bin= build configuration switch may be used to specify
1900 whether bin and sbin directories are merged, or if they should be
1901 included separately in $PATH and various listings of executable
1902 directories. The build configuration scripts will try to autodetect
1903 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
1904 system, but distributions are encouraged to configure this
1907 * A new -Dok-color= build configuration switch may be used to change
1908 the colour of "OK" status messages.
1910 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
1911 PrivateNetwork=yes was buggy in previous versions of systemd. This
1912 means that after the upgrade and daemon-reexec, any such units must
1915 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
1916 will not exclude read-only files owned by root from cleanup.
1918 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
1919 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
1920 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
1921 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
1922 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
1923 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
1924 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
1925 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
1926 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
1927 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
1928 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
1929 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
1930 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1931 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
1932 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
1933 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
1935 — Warsaw, 2018-03-05
1939 * Some keyboards come with a zoom see-saw or rocker which until now got
1940 mapped to the Linux "zoomin/out" keys in hwdb. However, these
1941 keycodes are not recognized by any major desktop. They now produce
1942 Up/Down key events so that they can be used for scrolling.
1944 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
1945 slightly: previously, if an argument was specified for lines of this
1946 type (i.e. the right-most column was set) this string was appended to
1947 existing files each time systemd-tmpfiles was run. This behaviour was
1948 different from what the documentation said, and not particularly
1949 useful, as repeated systemd-tmpfiles invocations would not be
1950 idempotent and grow such files without bounds. With this release
1951 behaviour has been altered to match what the documentation says:
1952 lines of this type only have an effect if the indicated files don't
1953 exist yet, and only then the argument string is written to the file.
1955 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
1956 systemd-tmpfiles behaviour: previously, read-only files owned by root
1957 were always excluded from the file "aging" algorithm (i.e. the
1958 automatic clean-up of directories like /tmp based on
1959 atime/mtime/ctime). We intend to drop this restriction, and age files
1960 by default even when owned by root and read-only. This behaviour was
1961 inherited from older tools, but there have been requests to remove
1962 it, and it's not obvious why this restriction was made in the first
1963 place. Please speak up now, if you are aware of software that reqires
1964 this behaviour, otherwise we'll remove the restriction in v238.
1966 * A new environment variable $SYSTEMD_OFFLINE is now understood by
1967 systemctl. It takes a boolean argument. If on, systemctl assumes it
1968 operates on an "offline" OS tree, and will not attempt to talk to the
1969 service manager. Previously, this mode was implicitly enabled if a
1970 chroot() environment was detected, and this new environment variable
1971 now provides explicit control.
1973 * .path and .socket units may now be created transiently, too.
1974 Previously only service, mount, automount and timer units were
1975 supported as transient units. The systemd-run tool has been updated
1976 to expose this new functionality, you may hence use it now to bind
1977 arbitrary commands to path or socket activation on-the-fly from the
1978 command line. Moreover, almost all properties are now exposed for the
1979 unit types that already supported transient operation.
1981 * The systemd-mount command gained support for a new --owner= parameter
1982 which takes a user name, which is then resolved and included in uid=
1983 and gid= mount options string of the file system to mount.
1985 * A new unit condition ConditionControlGroupController= has been added
1986 that checks whether a specific cgroup controller is available.
1988 * Unit files, udev's .link files, and systemd-networkd's .netdev and
1989 .network files all gained support for a new condition
1990 ConditionKernelVersion= for checking against specific kernel
1993 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
1994 support for configuring device flags in the Flags= setting. In the
1995 same files, the [Tunnel] section gained support for configuring
1996 AllowLocalRemote=. The [Route] section in .network files gained
1997 support for configuring InitialCongestionWindow=,
1998 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
1999 understands RapidCommit=.
2001 * systemd-networkd's DHCPv6 support gained support for Prefix
2004 * sd-bus gained support for a new "watch-bind" feature. When this
2005 feature is enabled, an sd_bus connection may be set up to connect to
2006 an AF_UNIX socket in the file system as soon as it is created. This
2007 functionality is useful for writing early-boot services that
2008 automatically connect to the system bus as soon as it is started,
2009 without ugly time-based polling. systemd-networkd and
2010 systemd-resolved have been updated to make use of this
2011 functionality. busctl exposes this functionality in a new
2012 --watch-bind= command line switch.
2014 * sd-bus will now optionally synthesize a local "Connected" signal as
2015 soon as a D-Bus connection is set up fully. This message mirrors the
2016 already existing "Disconnected" signal which is synthesized when the
2017 connection is terminated. This signal is generally useful but
2018 particularly handy in combination with the "watch-bind" feature
2019 described above. Synthesizing of this message has to be requested
2020 explicitly through the new API call sd_bus_set_connected_signal(). In
2021 addition a new call sd_bus_is_ready() has been added that checks
2022 whether a connection is fully set up (i.e. between the "Connected" and
2023 "Disconnected" signals).
2025 * sd-bus gained two new calls sd_bus_request_name_async() and
2026 sd_bus_release_name_async() for asynchronously registering bus
2027 names. Similar, there is now sd_bus_add_match_async() for installing
2028 a signal match asynchronously. All of systemd's own services have
2029 been updated to make use of these calls. Doing these operations
2030 asynchronously has two benefits: it reduces the risk of deadlocks in
2031 case of cyclic dependencies between bus services, and it speeds up
2032 service initialization since synchronization points for bus
2033 round-trips are removed.
2035 * sd-bus gained two new calls sd_bus_match_signal() and
2036 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
2037 and sd_bus_add_match_async() but instead of taking a D-Bus match
2038 string take match fields as normal function parameters.
2040 * sd-bus gained two new calls sd_bus_set_sender() and
2041 sd_bus_message_set_sender() for setting the sender name of outgoing
2042 messages (either for all outgoing messages or for just one specific
2043 one). These calls are only useful in direct connections as on
2044 brokered connections the broker fills in the sender anyway,
2045 overwriting whatever the client filled in.
2047 * sd-event gained a new pseudo-handle that may be specified on all API
2048 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
2049 used this refers to the default event loop object of the calling
2050 thread. Note however that this does not implicitly allocate one —
2051 which has to be done prior by using sd_event_default(). Similarly
2052 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
2053 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
2054 to the default bus of the specified type of the calling thread. Here
2055 too this does not implicitly allocate bus connection objects, this
2056 has to be done prior with sd_bus_default() and friends.
2058 * sd-event gained a new call pair
2059 sd_event_source_{get|set}_io_fd_own(). This may be used to request
2060 automatic closure of the file descriptor an IO event source watches
2061 when the event source is destroyed.
2063 * systemd-networkd gained support for natively configuring WireGuard
2066 * In previous versions systemd synthesized user records both for the
2067 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
2068 internally. In order to simplify distribution-wide renames of the
2069 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
2070 new transitional flag file has been added: if
2071 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
2072 user and group record within the systemd codebase is disabled.
2074 * systemd-notify gained a new --uid= option for selecting the source
2075 user/UID to use for notification messages sent to the service
2078 * journalctl gained a new --grep= option to list only entries in which
2079 the message matches a certain pattern. By default matching is case
2080 insensitive if the pattern is lowercase, and case sensitive
2081 otherwise. Option --case-sensitive=yes|no can be used to override
2082 this an specify case sensitivity or case insensitivity.
2084 * There's now a "systemd-analyze service-watchdogs" command for printing
2085 the current state of the service runtime watchdog, and optionally
2086 enabling or disabling the per-service watchdogs system-wide if given a
2087 boolean argument (i.e. the concept you configure in WatchdogSec=), for
2088 debugging purposes. There's also a kernel command line option
2089 systemd.service_watchdogs= for controlling the same.
2091 * Two new "log-level" and "log-target" options for systemd-analyze were
2092 added that merge the now deprecated get-log-level, set-log-level and
2093 get-log-target, set-log-target pairs. The deprecated options are still
2094 understood for backwards compatibility. The two new options print the
2095 current value when no arguments are given, and set them when a
2096 level/target is given as an argument.
2098 * sysusers.d's "u" lines now optionally accept both a UID and a GID
2099 specification, separated by a ":" character, in order to create users
2100 where UID and GID do not match.
2102 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
2103 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
2104 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
2105 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
2106 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
2107 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
2108 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
2109 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
2110 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
2111 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
2112 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
2113 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
2114 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
2115 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
2116 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
2117 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
2118 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
2119 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
2120 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
2121 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
2128 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
2129 in v235 has been extended to also set the dummy.ko module option
2130 numdummies=0, preventing the kernel from automatically creating
2131 dummy0. All dummy interfaces must now be explicitly created.
2133 * Unknown '%' specifiers in configuration files are now rejected. This
2134 applies to units and tmpfiles.d configuration. Any percent characters
2135 that are followed by a letter or digit that are not supposed to be
2136 interpreted as the beginning of a specifier should be escaped by
2137 doubling ("%%"). (So "size=5%" is still accepted, as well as
2138 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
2139 valid specifiers today.)
2141 * systemd-resolved now maintains a new dynamic
2142 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
2143 recommended to make /etc/resolv.conf a symlink to it. This file
2144 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
2145 includes dynamically acquired search domains, achieving more correct
2146 DNS resolution by software that bypasses local DNS APIs such as NSS.
2148 * The "uaccess" udev tag has been dropped from /dev/kvm and
2149 /dev/dri/renderD*. These devices now have the 0666 permissions by
2150 default (but this may be changed at build-time). /dev/dri/renderD*
2151 will now be owned by the "render" group along with /dev/kfd.
2153 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
2154 systemd-journal-gatewayd.service and
2155 systemd-journal-upload.service. This means "nss-systemd" must be
2156 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
2157 services are resolved properly.
2159 * In /etc/fstab two new mount options are now understood:
2160 x-systemd.makefs and x-systemd.growfs. The former has the effect that
2161 the configured file system is formatted before it is mounted, the
2162 latter that the file system is resized to the full block device size
2163 after it is mounted (i.e. if the file system is smaller than the
2164 partition it resides on, it's grown). This is similar to the fsck
2165 logic in /etc/fstab, and pulls in systemd-makefs@.service and
2166 systemd-growfs@.service as necessary, similar to
2167 systemd-fsck@.service. Resizing is currently only supported on ext4
2170 * In systemd-networkd, the IPv6 RA logic now optionally may announce
2171 DNS server and domain information.
2173 * Support for the LUKS2 on-disk format for encrypted partitions has
2174 been added. This requires libcryptsetup2 during compilation and
2177 * The systemd --user instance will now signal "readiness" when its
2178 basic.target unit has been reached, instead of when the run queue ran
2179 empty for the first time.
2181 * Tmpfiles.d with user configuration are now also supported.
2182 systemd-tmpfiles gained a new --user switch, and snippets placed in
2183 ~/.config/user-tmpfiles.d/ and corresponding directories will be
2184 executed by systemd-tmpfiles --user running in the new
2185 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
2186 running in the user session.
2188 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
2189 %S resolves to the top-level state directory (/var/lib for the system
2190 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
2191 top-level cache directory (/var/cache for the system instance,
2192 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
2193 logs directory (/var/log for the system instance,
2194 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
2195 existing %t specifier, that resolves to the top-level runtime
2196 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
2199 * journalctl learnt a new parameter --output-fields= for limiting the
2200 set of journal fields to output in verbose and JSON output modes.
2202 * systemd-timesyncd's configuration file gained a new option
2203 RootDistanceMaxSec= for setting the maximum root distance of servers
2204 it'll use, as well as the new options PollIntervalMinSec= and
2205 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
2207 * bootctl gained a new command "list" for listing all available boot
2208 menu items on systems that follow the boot loader specification.
2210 * systemctl gained a new --dry-run switch that shows what would be done
2211 instead of doing it, and is currently supported by the shutdown and
2214 * ConditionSecurity= can now detect the TOMOYO security module.
2216 * Unit file [Install] sections are now also respected in unit drop-in
2217 files. This is intended to be used by drop-ins under /usr/lib/.
2219 * systemd-firstboot may now also set the initial keyboard mapping.
2221 * Udev "changed" events for devices which are exposed as systemd
2222 .device units are now propagated to units specified in
2223 ReloadPropagatedFrom= as reload requests.
2225 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
2226 unit template name (i.e. a name in the form of 'foobar@.service',
2227 without the instance component between the '@' and - the '.'), then
2228 the escaped sysfs path of the device is automatically used as the
2231 * SystemCallFilter= in unit files has been extended so that an "errno"
2232 can be specified individually for each system call. Example:
2233 SystemCallFilter=~uname:EILSEQ.
2235 * The cgroup delegation logic has been substantially updated. Delegate=
2236 now optionally takes a list of controllers (instead of a boolean, as
2237 before), which lists the controllers to delegate at least.
2239 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
2241 * A new LogLevelMax= setting configures the maximum log level any
2242 process of the service may log at (i.e. anything with a lesser
2243 priority than what is specified is automatically dropped). A new
2244 LogExtraFields= setting allows configuration of additional journal
2245 fields to attach to all log records generated by any of the unit's
2248 * New StandardInputData= and StandardInputText= settings along with the
2249 new option StandardInput=data may be used to configure textual or
2250 binary data that shall be passed to the executed service process via
2251 standard input, encoded in-line in the unit file.
2253 * StandardInput=, StandardOutput= and StandardError= may now be used to
2254 connect stdin/stdout/stderr of executed processes directly with a
2255 file or AF_UNIX socket in the file system, using the new "file:" option.
2257 * A new unit file option CollectMode= has been added, that allows
2258 tweaking the garbage collection logic for units. It may be used to
2259 tell systemd to garbage collect units that have failed automatically
2260 (normally it only GCs units that exited successfully). systemd-run
2261 and systemd-mount expose this new functionality with a new -G option.
2263 * "machinectl bind" may now be used to bind mount non-directories
2264 (i.e. regularfiles, devices, fifos, sockets).
2266 * systemd-analyze gained a new verb "calendar" for validating and
2267 testing calendar time specifications to use for OnCalendar= in timer
2268 units. Besides validating the expression it will calculate the next
2269 time the specified expression would elapse.
2271 * In addition to the pre-existing FailureAction= unit file setting
2272 there's now SuccessAction=, for configuring a shutdown action to
2273 execute when a unit completes successfully. This is useful in
2274 particular inside containers that shall terminate after some workload
2275 has been completed. Also, both options are now supported for all unit
2276 types, not just services.
2278 * networkds's IP rule support gained two new options
2279 IncomingInterface= and OutgoingInterface= for configuring the incoming
2280 and outgoing interfaces of configured rules. systemd-networkd also
2281 gained support for "vxcan" network devices.
2283 * networkd gained a new setting RequiredForOnline=, taking a
2284 boolean. If set, systemd-wait-online will take it into consideration
2285 when determining that the system is up, otherwise it will ignore the
2286 interface for this purpose.
2288 * The sd_notify() protocol gained support for a new operation: with
2289 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
2290 store again, ahead of POLLHUP or POLLERR when they are removed
2293 * A new document doc/UIDS-GIDS.md has been added to the source tree,
2294 that documents the UID/GID range and assignment assumptions and
2295 requirements of systemd.
2297 * The watchdog device PID 1 will ping may now be configured through the
2298 WatchdogDevice= configuration file setting, or by setting the
2299 systemd.watchdog_service= kernel commandline option.
2301 * systemd-resolved's gained support for registering DNS-SD services on
2302 the local network using MulticastDNS. Services may either be
2303 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
2304 the same dir below /run, /usr/lib), or through its D-Bus API.
2306 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
2307 extend the effective start, runtime, and stop time. The service must
2308 continue to send EXTEND_TIMEOUT_USEC within the period specified to
2309 prevent the service manager from making the service as timedout.
2311 * systemd-resolved's DNSSEC support gained support for RFC 8080
2312 (Ed25519 keys and signatures).
2314 * The systemd-resolve command line tool gained a new set of options
2315 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
2316 --set-nta= and --revert to configure per-interface DNS configuration
2317 dynamically during runtime. It's useful for pushing DNS information
2318 into systemd-resolved from DNS hook scripts that various interface
2319 managing software supports (such as pppd).
2321 * systemd-nspawn gained a new --network-namespace-path= command line
2322 option, which may be used to make a container join an existing
2323 network namespace, by specifying a path to a "netns" file.
2325 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
2326 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
2327 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
2328 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
2329 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
2330 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
2331 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
2332 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
2333 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
2334 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
2335 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
2336 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
2337 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
2338 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
2339 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
2340 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
2341 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
2342 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
2343 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
2344 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
2345 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
2346 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
2347 Jędrzejewski-Szmek, Zeal Jagannatha
2349 — Berlin, 2017-12-14
2353 * INCOMPATIBILITY: systemd-logind.service and other long-running
2354 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
2355 communication with the outside. This generally improves security of
2356 the system, and is in almost all cases a safe and good choice, as
2357 these services do not and should not provide any network-facing
2358 functionality. However, systemd-logind uses the glibc NSS API to
2359 query the user database. This creates problems on systems where NSS
2360 is set up to directly consult network services for user database
2361 lookups. In particular, this creates incompatibilities with the
2362 "nss-nis" module, which attempts to directly contact the NIS/YP
2363 network servers it is configured for, and will now consistently
2364 fail. In such cases, it is possible to turn off IP sandboxing for
2365 systemd-logind.service (set IPAddressDeny= in its [Service] section
2366 to the empty string, via a .d/ unit file drop-in). Downstream
2367 distributions might want to update their nss-nis packaging to include
2368 such a drop-in snippet, accordingly, to hide this incompatibility
2369 from the user. Another option is to make use of glibc's nscd service
2370 to proxy such network requests through a privilege-separated, minimal
2371 local caching daemon, or to switch to more modern technologies such
2372 sssd, whose NSS hook-ups generally do not involve direct network
2373 access. In general, we think it's definitely time to question the
2374 implementation choices of nss-nis, i.e. whether it's a good idea
2375 today to embed a network-facing loadable module into all local
2376 processes that need to query the user database, including the most
2377 trivial and benign ones, such as "ls". For more details about
2378 IPAddressDeny= see below.
2380 * A new modprobe.d drop-in is now shipped by default that sets the
2381 bonding module option max_bonds=0. This overrides the kernel default,
2382 to avoid conflicts and ambiguity as to whether or not bond0 should be
2383 managed by systemd-networkd or not. This resolves multiple issues
2384 with bond0 properties not being applied, when bond0 is configured
2385 with systemd-networkd. Distributors may choose to not package this,
2386 however in that case users will be prevented from correctly managing
2387 bond0 interface using systemd-networkd.
2389 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
2390 which print the logging level and target of the system manager. They
2391 complement the existing "set-log-level" and "set-log-target" verbs
2392 used to change those values.
2394 * journald.conf gained a new boolean setting ReadKMsg= which defaults
2395 to on. If turned off kernel log messages will not be read by
2396 systemd-journald or included in the logs. It also gained a new
2397 setting LineMax= for configuring the maximum line length in
2398 STDOUT/STDERR log streams. The new default for this value is 48K, up
2399 from the previous hardcoded 2048.
2401 * A new unit setting RuntimeDirectoryPreserve= has been added, which
2402 allows more detailed control of what to do with a runtime directory
2403 configured with RuntimeDirectory= (i.e. a directory below /run or
2404 $XDG_RUNTIME_DIR) after a unit is stopped.
2406 * The RuntimeDirectory= setting for units gained support for creating
2407 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
2408 one top-level directory.
2410 * Units gained new options StateDirectory=, CacheDirectory=,
2411 LogsDirectory= and ConfigurationDirectory= which are closely related
2412 to RuntimeDirectory= but manage per-service directories below
2413 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
2414 possible to write unit files which when activated automatically gain
2415 properly owned service specific directories in these locations, thus
2416 making unit files self-contained and increasing compatibility with
2417 stateless systems and factory reset where /etc or /var are
2418 unpopulated at boot. Matching these new settings there's also
2419 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
2420 ConfigurationDirectoryMode= for configuring the access mode of these
2421 directories. These settings are particularly useful in combination
2422 with DynamicUser=yes as they provide secure, properly-owned,
2423 writable, and stateful locations for storage, excluded from the
2424 sandbox that such services live in otherwise.
2426 * Automake support has been removed from this release. systemd is now
2429 * systemd-journald will now aggressively cache client metadata during
2430 runtime, speeding up log write performance under pressure. This comes
2431 at a small price though: as much of the metadata is read
2432 asynchronously from /proc/ (and isn't implicitly attached to log
2433 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
2434 metadata stored alongside a log entry might be slightly
2435 out-of-date. Previously it could only be slightly newer than the log
2436 message. The time window is small however, and given that the kernel
2437 is unlikely to be improved anytime soon in this regard, this appears
2440 * nss-myhostname/systemd-resolved will now by default synthesize an
2441 A/AAAA resource record for the "_gateway" hostname, pointing to the
2442 current default IP gateway. Previously it did that for the "gateway"
2443 name, hampering adoption, as some distributions wanted to leave that
2444 host name open for local use. The old behaviour may still be
2445 requested at build time.
2447 * systemd-networkd's [Address] section in .network files gained a new
2448 Scope= setting for configuring the IP address scope. The [Network]
2449 section gained a new boolean setting ConfigureWithoutCarrier= that
2450 tells systemd-networkd to ignore link sensing when configuring the
2451 device. The [DHCP] section gained a new Anonymize= boolean option for
2452 turning on a number of options suggested in RFC 7844. A new
2453 [RoutingPolicyRule] section has been added for configuring the IP
2454 routing policy. The [Route] section has gained support for a new
2455 Type= setting which permits configuring
2456 blackhole/unreachable/prohibit routes.
2458 * The [VRF] section in .netdev files gained a new Table= setting for
2459 configuring the routing table to use. The [Tunnel] section gained a
2460 new Independent= boolean field for configuring tunnels independent of
2461 an underlying network interface. The [Bridge] section gained a new
2462 GroupForwardMask= option for configuration of propagation of link
2463 local frames between bridge ports.
2465 * The WakeOnLan= setting in .link files gained support for a number of
2466 new modes. A new TCP6SegmentationOffload= setting has been added for
2467 configuring TCP/IPv6 hardware segmentation offload.
2469 * The IPv6 RA sender implementation may now optionally send out RDNSS
2470 and RDNSSL records to supply DNS configuration to peers.
2472 * systemd-nspawn gained support for a new --system-call-filter= command
2473 line option for adding and removing entries in the default system
2474 call filter it applies. Moreover systemd-nspawn has been changed to
2475 implement a system call whitelist instead of a blacklist.
2477 * systemd-run gained support for a new --pipe command line option. If
2478 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
2479 are directly passed on to the activated transient service
2480 executable. This allows invoking arbitrary processes as systemd
2481 services (for example to take benefit of dependency management,
2482 accounting management, resource management or log management that is
2483 done automatically for services) — while still allowing them to be
2484 integrated in a classic UNIX shell pipeline.
2486 * When a service sends RELOAD=1 via sd_notify() and reload propagation
2487 using ReloadPropagationTo= is configured, a reload is now propagated
2488 to configured units. (Previously this was only done on explicitly
2489 requested reloads, using "systemctl reload" or an equivalent
2492 * For each service unit a restart counter is now kept: it is increased
2493 each time the service is restarted due to Restart=, and may be
2494 queried using "systemctl show -p NRestarts …".
2496 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
2497 @signal and @timer have been added, for usage with SystemCallFilter=
2498 in unit files and the new --system-call-filter= command line option
2499 of systemd-nspawn (see above).
2501 * ExecStart= lines in unit files gained two new modifiers: when a
2502 command line is prefixed with "!" the command will be executed as
2503 configured, except for the credentials applied by
2504 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
2505 "+", but does still apply namespacing options unlike "+". There's
2506 also "!!" now, which is mostly identical, but becomes a NOP on
2507 systems that support ambient capabilities. This is useful to write
2508 unit files that work with ambient capabilities where possible but
2509 automatically fall back to traditional privilege dropping mechanisms
2510 on systems where this is not supported.
2512 * ListenNetlink= settings in socket units now support RDMA netlink
2515 * A new unit file setting LockPersonality= has been added which permits
2516 locking down the chosen execution domain ("personality") of a service
2519 * A new special target "getty-pre.target" has been added, which is
2520 ordered before all text logins, and may be used to order services
2521 before textual logins acquire access to the console.
2523 * systemd will now attempt to load the virtio-rng.ko kernel module very
2524 early on if a VM environment supporting this is detected. This should
2525 improve entropy during early boot in virtualized environments.
2527 * A _netdev option is now supported in /etc/crypttab that operates in a
2528 similar way as the same option in /etc/fstab: it permits configuring
2529 encrypted devices that need to be ordered after the network is up.
2530 Following this logic, two new special targets
2531 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
2532 added that are to cryptsetup.target what remote-fs.target and
2533 remote-fs-pre.target are to local-fs.target.
2535 * Service units gained a new UnsetEnvironment= setting which permits
2536 unsetting specific environment variables for services that are
2537 normally passed to it (for example in order to mask out locale
2538 settings for specific services that can't deal with it).
2540 * Units acquired a new boolean option IPAccounting=. When turned on, IP
2541 traffic accounting (packet count as well as byte count) is done for
2542 the service, and shown as part of "systemctl status" or "systemd-run
2545 * Service units acquired two new options IPAddressAllow= and
2546 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
2547 for configuring a simple IP access control list for all sockets of
2548 the unit. These options are available also on .slice and .socket
2549 units, permitting flexible access list configuration for individual
2550 services as well as groups of services (as defined by a slice unit),
2551 including system-wide. Note that IP ACLs configured this way are
2552 enforced on every single IPv4 and IPv6 socket created by any process
2553 of the service unit, and apply to ingress as well as egress traffic.
2555 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
2556 structured log message is generated each time the unit is stopped,
2557 containing information about the consumed resources of this
2560 * A new setting KeyringMode= has been added to unit files, which may be
2561 used to control how the kernel keyring is set up for executed
2564 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
2565 "systemctl kexec" and "systemctl exit" are now always asynchronous in
2566 behaviour (that is: these commands return immediately after the
2567 operation was enqueued instead of waiting for the operation to
2568 complete). Previously, "systemctl poweroff" and "systemctl reboot"
2569 were asynchronous on systems using systemd-logind (i.e. almost
2570 always, and like they were on sysvinit), and the other three commands
2571 were unconditionally synchronous. With this release this is cleaned
2572 up, and callers will see the same asynchronous behaviour on all
2573 systems for all five operations.
2575 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
2578 * .timer units now accept calendar specifications in other timezones
2579 than UTC or the local timezone.
2581 * The tmpfiles snippet var.conf has been changed to create
2582 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
2583 the "utmp" group already, and it appears to be generally understood
2584 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
2585 databases. Previously this was implemented correctly for all these
2586 databases excepts btmp, which has been opened up like this now
2587 too. Note that while the other databases are world-readable
2588 (i.e. 0644), btmp is not and remains more restrictive.
2590 * The systemd-resolve tool gained a new --reset-server-features
2591 switch. When invoked like this systemd-resolved will forget
2592 everything it learnt about the features supported by the configured
2593 upstream DNS servers, and restarts the feature probing logic on the
2594 next resolver look-up for them at the highest feature level
2597 * The status dump systemd-resolved sends to the logs upon receiving
2598 SIGUSR1 now also includes information about all DNS servers it is
2599 configured to use, and the features levels it probed for them.
2601 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
2602 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
2603 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
2604 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
2605 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
2606 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
2607 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
2608 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
2609 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
2610 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
2611 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
2612 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
2613 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
2614 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
2615 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
2616 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
2617 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
2618 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
2619 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
2620 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2622 — Berlin, 2017-10-06
2626 * Meson is now supported as build system in addition to Automake. It is
2627 our plan to remove Automake in one of our next releases, so that
2628 Meson becomes our exclusive build system. Hence, please start using
2629 the Meson build system in your downstream packaging. There's plenty
2630 of documentation around how to use Meson, the extremely brief
2633 ./autogen.sh && ./configure && make && sudo make install
2637 meson build && ninja -C build && sudo ninja -C build install
2639 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
2640 which permits configuring a timeout on the time a job is
2641 running. This is particularly useful for setting timeouts on jobs for
2644 * Unit files gained two new options ConditionUser= and ConditionGroup=
2645 for conditionalizing units based on the identity of the user/group
2646 running a systemd user instance.
2648 * systemd-networkd now understands a new FlowLabel= setting in the
2649 [VXLAN] section of .network files, as well as a Priority= in
2650 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
2651 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
2652 gained support for configuration of GENEVE links, and IPv6 address
2653 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
2655 * .link files now understand a new Port= setting.
2657 * systemd-networkd's DHCP support gained support for DHCP option 119
2658 (domain search list).
2660 * systemd-networkd gained support for serving IPv6 address ranges using
2661 the Router Advertisement protocol. The new .network configuration
2662 section [IPv6Prefix] may be used to configure the ranges to
2663 serve. This is implemented based on a new, minimal, native server
2664 implementation of RA.
2666 * journalctl's --output= switch gained support for a new parameter
2667 "short-iso-precise" for a mode where timestamps are shown as precise
2670 * systemd-udevd's "net_id" builtin may now generate stable network
2671 interface names from IBM PowerVM VIO devices as well as ACPI platform
2674 * MulticastDNS support in systemd-resolved may now be explicitly
2675 enabled/disabled using the new MulticastDNS= configuration file
2678 * systemd-resolved may now optionally use libidn2 instead of the libidn
2679 for processing internationalized domain names. Support for libidn2
2680 should be considered experimental and should not be enabled by
2683 * "machinectl pull-tar" and related call may now do verification of
2684 downloaded images using SUSE-style .sha256 checksum files in addition
2685 to the already existing support for validating using Ubuntu-style
2688 * sd-bus gained support for a new sd_bus_message_appendv() call which
2689 is va_list equivalent of sd_bus_message_append().
2691 * sd-boot gained support for validating images using SHIM/MOK.
2693 * The SMACK code learnt support for "onlycap".
2695 * systemd-mount --umount is now much smarter in figuring out how to
2696 properly unmount a device given its mount or device path.
2698 * The code to call libnss_dns as a fallback from libnss_resolve when
2699 the communication with systemd-resolved fails was removed. This
2700 fallback was redundant and interfered with the [!UNAVAIL=return]
2701 suffix. See nss-resolve(8) for the recommended configuration.
2703 * systemd-logind may now be restarted without losing state. It stores
2704 the file descriptors for devices it manages in the system manager
2705 using the FDSTORE= mechanism. Please note that further changes in
2706 other components may be required to make use of this (for example
2707 Xorg has code to listen for stops of systemd-logind and terminate
2708 itself when logind is stopped or restarted, in order to avoid using
2709 stale file descriptors for graphical devices, which is now
2710 counterproductive and must be reverted in order for restarts of
2711 systemd-logind to be safe. See
2712 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
2714 * All kernel-install plugins are called with the environment variable
2715 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
2716 /etc/machine-id. If the machine ID could not be determined,
2717 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
2718 anything in the entry directory (passed as the second argument) if
2719 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
2720 temporary directory is passed as the entry directory and removed
2721 after all the plugins exit.
2723 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
2724 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
2725 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
2726 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
2727 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
2728 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
2729 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
2730 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
2731 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
2732 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
2733 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
2734 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
2735 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
2736 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
2737 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
2738 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
2739 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
2740 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
2741 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
2742 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
2743 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
2744 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
2745 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
2746 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
2747 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
2748 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
2749 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
2750 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
2751 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
2754 — Berlin, 2017-07-12
2758 * The "hybrid" control group mode has been modified to improve
2759 compatibility with "legacy" cgroups-v1 setups. Specifically, the
2760 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
2761 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
2762 cgroups-v1 hierarchy), the only externally visible change being that
2763 the cgroups-v2 hierarchy is also mounted, to
2764 /sys/fs/cgroup/unified. This should provide a large degree of
2765 compatibility with "legacy" cgroups-v1, while taking benefit of the
2766 better management capabilities of cgroups-v2.
2768 * The default control group setup mode may be selected both a boot-time
2769 via a set of kernel command line parameters (specifically:
2770 systemd.unified_cgroup_hierarchy= and
2771 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
2772 default selected on the configure command line
2773 (--with-default-hierarchy=). The upstream default is "hybrid"
2774 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
2775 this will change in a future systemd version to be "unified" (pure
2776 cgroups-v2 mode). The third option for the compile time option is
2777 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
2778 distributions to default to "hybrid" mode for release distributions,
2779 starting with v233. We recommend "unified" for development
2780 distributions (specifically: distributions such as Fedora's rawhide)
2781 as that's where things are headed in the long run. Use "legacy" for
2782 greatest stability and compatibility only.
2784 * Note one current limitation of "unified" and "hybrid" control group
2785 setup modes: the kernel currently does not permit the systemd --user
2786 instance (i.e. unprivileged code) to migrate processes between two
2787 disconnected cgroup subtrees, even if both are managed and owned by
2788 the user. This effectively means "systemd-run --user --scope" doesn't
2789 work when invoked from outside of any "systemd --user" service or
2790 scope. Specifically, it is not supported from session scopes. We are
2791 working on fixing this in a future systemd version. (See #3388 for
2792 further details about this.)
2794 * DBus policy files are now installed into /usr rather than /etc. Make
2795 sure your system has dbus >= 1.9.18 running before upgrading to this
2796 version, or override the install path with --with-dbuspolicydir= .
2798 * All python scripts shipped with systemd (specifically: the various
2799 tests written in Python) now require Python 3.
2801 * systemd unit tests can now run standalone (without the source or
2802 build directories), and can be installed into /usr/lib/systemd/tests/
2803 with 'make install-tests'.
2805 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
2806 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
2809 * Support for the %c, %r, %R specifiers in unit files has been
2810 removed. Specifiers are not supposed to be dependent on configuration
2811 in the unit file itself (so that they resolve the same regardless
2812 where used in the unit files), but these specifiers were influenced
2813 by the Slice= option.
2815 * The shell invoked by debug-shell.service now defaults to /bin/sh in
2816 all cases. If distributions want to use a different shell for this
2817 purpose (for example Fedora's /sbin/sushell) they need to specify
2818 this explicitly at configure time using --with-debug-shell=.
2820 * The confirmation spawn prompt has been reworked to offer the
2823 (c)ontinue, proceed without asking anymore
2824 (D)ump, show the state of the unit
2825 (f)ail, don't execute the command and pretend it failed
2827 (i)nfo, show a short summary of the unit
2828 (j)obs, show jobs that are in progress
2829 (s)kip, don't execute the command and pretend it succeeded
2830 (y)es, execute the command
2832 The 'n' choice for the confirmation spawn prompt has been removed,
2833 because its meaning was confusing.
2835 The prompt may now also be redirected to an alternative console by
2836 specifying the console as parameter to systemd.confirm_spawn=.
2838 * Services of Type=notify require a READY=1 notification to be sent
2839 during startup. If no such message is sent, the service now fails,
2840 even if the main process exited with a successful exit code.
2842 * Services that fail to start up correctly now always have their
2843 ExecStopPost= commands executed. Previously, they'd enter "failed"
2844 state directly, without executing these commands.
2846 * The option MulticastDNS= of network configuration files has acquired
2847 an actual implementation. With MulticastDNS=yes a host can resolve
2848 names of remote hosts and reply to mDNS A and AAAA requests.
2850 * When units are about to be started an additional check is now done to
2851 ensure that all dependencies of type BindsTo= (when used in
2852 combination with After=) have been started.
2854 * systemd-analyze gained a new verb "syscall-filter" which shows which
2855 system call groups are defined for the SystemCallFilter= unit file
2856 setting, and which system calls they contain.
2858 * A new system call filter group "@filesystem" has been added,
2859 consisting of various file system related system calls. Group
2860 "@reboot" has been added, covering reboot, kexec and shutdown related
2861 calls. Finally, group "@swap" has been added covering swap
2862 configuration related calls.
2864 * A new unit file option RestrictNamespaces= has been added that may be
2865 used to restrict access to the various process namespace types the
2866 Linux kernel provides. Specifically, it may be used to take away the
2867 right for a service unit to create additional file system, network,
2868 user, and other namespaces. This sandboxing option is particularly
2869 relevant due to the high amount of recently discovered namespacing
2870 related vulnerabilities in the kernel.
2872 * systemd-udev's .link files gained support for a new AutoNegotiation=
2873 setting for configuring Ethernet auto-negotiation.
2875 * systemd-networkd's .network files gained support for a new
2876 ListenPort= setting in the [DHCP] section to explicitly configure the
2877 UDP client port the DHCP client shall listen on.
2879 * .network files gained a new Unmanaged= boolean setting for explicitly
2880 excluding one or more interfaces from management by systemd-networkd.
2882 * The systemd-networkd ProxyARP= option has been renamed to
2883 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
2884 renamed to ReduceARPProxy=. The old names continue to be available
2887 * systemd-networkd gained support for configuring IPv6 Proxy NDP
2888 addresses via the new IPv6ProxyNDPAddress= .network file setting.
2890 * systemd-networkd's bonding device support gained support for two new
2891 configuration options ActiveSlave= and PrimarySlave=.
2893 * The various options in the [Match] section of .network files gained
2894 support for negative matching.
2896 * New systemd-specific mount options are now understood in /etc/fstab:
2898 x-systemd.mount-timeout= may be used to configure the maximum
2899 permitted runtime of the mount command.
2901 x-systemd.device-bound may be set to bind a mount point to its
2902 backing device unit, in order to automatically remove a mount point
2903 if its backing device is unplugged. This option may also be
2904 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
2905 on the block device, which is now automatically set for all CDROM
2906 drives, so that mounted CDs are automatically unmounted when they are
2907 removed from the drive.
2909 x-systemd.after= and x-systemd.before= may be used to explicitly
2910 order a mount after or before another unit or mount point.
2912 * Enqueued start jobs for device units are now automatically garbage
2913 collected if there are no jobs waiting for them anymore.
2915 * systemctl list-jobs gained two new switches: with --after, for every
2916 queued job the jobs it's waiting for are shown; with --before the
2917 jobs which it's blocking are shown.
2919 * systemd-nspawn gained support for ephemeral boots from disk images
2920 (or in other words: --ephemeral and --image= may now be
2921 combined). Moreover, ephemeral boots are now supported for normal
2922 directories, even if the backing file system is not btrfs. Of course,
2923 if the file system does not support file system snapshots or
2924 reflinks, the initial copy operation will be relatively expensive, but
2925 this should still be suitable for many use cases.
2927 * Calendar time specifications in .timer units now support
2928 specifications relative to the end of a month by using "~" instead of
2929 "-" as separator between month and day. For example, "*-02~03" means
2930 "the third last day in February". In addition a new syntax for
2931 repeated events has been added using the "/" character. For example,
2932 "9..17/2:00" means "every two hours from 9am to 5pm".
2934 * systemd-socket-proxyd gained a new parameter --connections-max= for
2935 configuring the maximum number of concurrent connections.
2937 * sd-id128 gained a new API for generating unique IDs for the host in a
2938 way that does not leak the machine ID. Specifically,
2939 sd_id128_get_machine_app_specific() derives an ID based on the
2940 machine ID a in well-defined, non-reversible, stable way. This is
2941 useful whenever an identifier for the host is needed but where the
2942 identifier shall not be useful to identify the system beyond the
2943 scope of the application itself. (Internally this uses HMAC-SHA256 as
2944 keyed hash function using the machine ID as input.)
2946 * NotifyAccess= gained a new supported value "exec". When set
2947 notifications are accepted from all processes systemd itself invoked,
2948 including all control processes.
2950 * .nspawn files gained support for defining overlay mounts using the
2951 Overlay= and OverlayReadOnly= options. Previously this functionality
2952 was only available on the systemd-nspawn command line.
2954 * systemd-nspawn's --bind= and --overlay= options gained support for
2955 bind/overlay mounts whose source lies within the container tree by
2956 prefixing the source path with "+".
2958 * systemd-nspawn's --bind= and --overlay= options gained support for
2959 automatically allocating a temporary source directory in /var/tmp
2960 that is removed when the container dies. Specifically, if the source
2961 directory is specified as empty string this mechanism is selected. An
2962 example usage is --overlay=+/var::/var, which creates an overlay
2963 mount based on the original /var contained in the image, overlayed
2964 with a temporary directory in the host's /var/tmp. This way changes
2965 to /var are automatically flushed when the container shuts down.
2967 * systemd-nspawn --image= option does now permit raw file system block
2968 devices (in addition to images containing partition tables, as
2971 * The disk image dissection logic in systemd-nspawn gained support for
2972 automatically setting up LUKS encrypted as well as Verity protected
2973 partitions. When a container is booted from an encrypted image the
2974 passphrase is queried at start-up time. When a container with Verity
2975 data is started, the root hash is search in a ".roothash" file
2976 accompanying the disk image (alternatively, pass the root hash via
2977 the new --root-hash= command line option).
2979 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
2980 be used to dissect disk images the same way as systemd-nspawn does
2981 it, following the Bootable Partition Specification. It may even be
2982 used to mount disk images with complex partition setups (including
2983 LUKS and Verity partitions) to a local host directory, in order to
2984 inspect them. This tool is not considered public API (yet), and is
2985 thus not installed into /usr/bin. Please do not rely on its
2986 existence, since it might go away or be changed in later systemd
2989 * A new generator "systemd-verity-generator" has been added, similar in
2990 style to "systemd-cryptsetup-generator", permitting automatic setup of
2991 Verity root partitions when systemd boots up. In order to make use of
2992 this your partition setup should follow the Discoverable Partitions
2993 Specification, and the GPT partition ID of the root file system
2994 partition should be identical to the upper 128bit of the Verity root
2995 hash. The GPT partition ID of the Verity partition protecting it
2996 should be the lower 128bit of the Verity root hash. If the partition
2997 image follows this model it is sufficient to specify a single
2998 "roothash=" kernel command line argument to both configure which root
2999 image and verity partition to use as well as the root hash for
3000 it. Note that systemd-nspawn's Verity support follows the same
3001 semantics, meaning that disk images with proper Verity data in place
3002 may be booted in containers with systemd-nspawn as well as on
3003 physical systems via the verity generator. Also note that the "mkosi"
3004 tool available at https://github.com/systemd/mkosi has been updated
3005 to generate Verity protected disk images following this scheme. In
3006 fact, it has been updated to generate disk images that optionally
3007 implement a complete UEFI SecureBoot trust chain, involving a signed
3008 kernel and initrd image that incorporates such a root hash as well as
3009 a Verity-enabled root partition.
3011 * The hardware database (hwdb) udev supports has been updated to carry
3012 accelerometer quirks.
3014 * All system services are now run with a fresh kernel keyring set up
3015 for them. The invocation ID is stored by default in it, thus
3016 providing a safe, non-overridable way to determine the invocation
3019 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
3020 options for bind mounting arbitrary paths in a service-specific
3021 way. When these options are used, arbitrary host or service files and
3022 directories may be mounted to arbitrary locations in the service's
3025 * Documentation has been added that lists all of systemd's low-level
3026 environment variables:
3028 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
3030 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
3031 whether a specific socket file descriptor matches a specified socket
3034 * systemd-firstboot has been updated to check for the
3035 systemd.firstboot= kernel command line option. It accepts a boolean
3036 and when set to false the first boot questions are skipped.
3038 * systemd-fstab-generator has been updated to check for the
3039 systemd.volatile= kernel command line option, which either takes an
3040 optional boolean parameter or the special value "state". If used the
3041 system may be booted in a "volatile" boot mode. Specifically,
3042 "systemd.volatile" is used, the root directory will be mounted as
3043 tmpfs, and only /usr is mounted from the actual root file system. If
3044 "systemd.volatile=state" is used, the root directory will be mounted
3045 as usual, but /var is mounted as tmpfs. This concept provides similar
3046 functionality as systemd-nspawn's --volatile= option, but provides it
3047 on physical boots. Use this option for implementing stateless
3048 systems, or testing systems with all state and/or configuration reset
3049 to the defaults. (Note though that many distributions are not
3050 prepared to boot up without a populated /etc or /var, though.)
3052 * systemd-gpt-auto-generator gained support for LUKS encrypted root
3053 partitions. Previously it only supported LUKS encrypted partitions
3054 for all other uses, except for the root partition itself.
3056 * Socket units gained support for listening on AF_VSOCK sockets for
3057 communication in virtualized QEMU environments.
3059 * The "configure" script gained a new option --with-fallback-hostname=
3060 for specifying the fallback hostname to use if none is configured in
3061 /etc/hostname. For example, by specifying
3062 --with-fallback-hostname=fedora it is possible to default to a
3063 hostname of "fedora" on pristine installations.
3065 * systemd-cgls gained support for a new --unit= switch for listing only
3066 the control groups of a specific unit. Similar --user-unit= has been
3067 added for listing only the control groups of a specific user unit.
3069 * systemd-mount gained a new --umount switch for unmounting a mount or
3070 automount point (and all mount/automount points below it).
3072 * systemd will now refuse full configuration reloads (via systemctl
3073 daemon-reload and related calls) unless at least 16MiB of free space
3074 are available in /run. This is a safety precaution in order to ensure
3075 that generators can safely operate after the reload completed.
3077 * A new unit file option RootImage= has been added, which has a similar
3078 effect as RootDirectory= but mounts the service's root directory from
3079 a disk image instead of plain directory. This logic reuses the same
3080 image dissection and mount logic that systemd-nspawn already uses,
3081 and hence supports any disk images systemd-nspawn supports, including
3082 those following the Discoverable Partition Specification, as well as
3083 Verity enabled images. This option enables systemd to run system
3084 services directly off disk images acting as resource bundles,
3085 possibly even including full integrity data.
3087 * A new MountAPIVFS= unit file option has been added, taking a boolean
3088 argument. If enabled /proc, /sys and /dev (collectively called the
3089 "API VFS") will be mounted for the service. This is only relevant if
3090 RootDirectory= or RootImage= is used for the service, as these mounts
3091 are of course in place in the host mount namespace anyway.
3093 * systemd-nspawn gained support for a new --pivot-root= switch. If
3094 specified the root directory within the container image is pivoted to
3095 the specified mount point, while the original root disk is moved to a
3096 different place. This option enables booting of ostree images
3097 directly with systemd-nspawn.
3099 * The systemd build scripts will no longer complain if the NTP server
3100 addresses are not changed from the defaults. Google now supports
3101 these NTP servers officially. We still recommend downstreams to
3102 properly register an NTP pool with the NTP pool project though.
3104 * coredumpctl gained a new "--reverse" option for printing the list
3105 of coredumps in reverse order.
3107 * coredumpctl will now show additional information about truncated and
3108 inaccessible coredumps, as well as coredumps that are still being
3109 processed. It also gained a new --quiet switch for suppressing
3110 additional informational message in its output.
3112 * coredumpctl gained support for only showing coredumps newer and/or
3113 older than specific timestamps, using the new --since= and --until=
3114 options, reminiscent of journalctl's options by the same name.
3116 * The systemd-coredump logic has been improved so that it may be reused
3117 to collect backtraces in non-compiled languages, for example in
3118 scripting languages such as Python.
3120 * machinectl will now show the UID shift of local containers, if user
3121 namespacing is enabled for them.
3123 * systemd will now optionally run "environment generator" binaries at
3124 configuration load time. They may be used to add environment
3125 variables to the environment block passed to services invoked. One
3126 user environment generator is shipped by default that sets up
3127 environment variables based on files dropped into /etc/environment.d
3128 and ~/.config/environment.d/.
3130 * systemd-resolved now includes the new, recently published 2017 DNSSEC
3133 * hostnamed has been updated to report a new chassis type of
3134 "convertible" to cover "foldable" laptops that can both act as a
3135 tablet and as a laptop, such as various Lenovo Yoga devices.
3137 Contributions from: Adrián López, Alexander Galanin, Alexander
3138 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
3139 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
3140 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
3141 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
3142 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
3143 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
3144 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
3145 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
3146 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
3147 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
3148 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
3149 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
3150 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
3151 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
3152 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
3153 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
3154 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
3155 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
3156 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
3157 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
3158 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
3159 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
3160 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
3161 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
3162 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
3163 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
3164 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
3167 — Berlin, 2017-03-01
3171 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
3172 RestrictAddressFamilies= enabled. These sandboxing options should
3173 generally be compatible with the various external udev call-out
3174 binaries we are aware of, however there may be exceptions, in
3175 particular when exotic languages for these call-outs are used. In
3176 this case, consider turning off these settings locally.
3178 * The new RemoveIPC= option can be used to remove IPC objects owned by
3179 the user or group of a service when that service exits.
3181 * The new ProtectKernelModules= option can be used to disable explicit
3182 load and unload operations of kernel modules by a service. In
3183 addition access to /usr/lib/modules is removed if this option is set.
3185 * ProtectSystem= option gained a new value "strict", which causes the
3186 whole file system tree with the exception of /dev, /proc, and /sys,
3187 to be remounted read-only for a service.
3189 * The new ProtectKernelTunables= option can be used to disable
3190 modification of configuration files in /sys and /proc by a service.
3191 Various directories and files are remounted read-only, so access is
3192 restricted even if the file permissions would allow it.
3194 * The new ProtectControlGroups= option can be used to disable write
3195 access by a service to /sys/fs/cgroup.
3197 * Various systemd services have been hardened with
3198 ProtectKernelTunables=yes, ProtectControlGroups=yes,
3199 RestrictAddressFamilies=.
3201 * Support for dynamically creating users for the lifetime of a service
3202 has been added. If DynamicUser=yes is specified, user and group IDs
3203 will be allocated from the range 61184..65519 for the lifetime of the
3204 service. They can be resolved using the new nss-systemd.so NSS
3205 module. The module must be enabled in /etc/nsswitch.conf. Services
3206 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
3207 any resources allocated by the service will be cleaned up when the
3208 service exits. They also have ProtectHome=read-only and
3209 ProtectSystem=strict enabled, so they are not able to make any
3210 permanent modifications to the system.
3212 * The nss-systemd module also always resolves root and nobody, making
3213 it possible to have no /etc/passwd or /etc/group files in minimal
3214 container or chroot environments.
3216 * Services may be started with their own user namespace using the new
3217 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
3218 under which the service is running are mapped. All other users are
3221 * Support for the cgroup namespace has been added to systemd-nspawn. If
3222 supported by kernel, the container system started by systemd-nspawn
3223 will have its own view of the cgroup hierarchy. This new behaviour
3224 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
3226 * The new MemorySwapMax= option can be used to limit the maximum swap
3227 usage under the unified cgroup hierarchy.
3229 * Support for the CPU controller in the unified cgroup hierarchy has
3230 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
3231 options. This controller requires out-of-tree patches for the kernel
3232 and the support is provisional.
3234 * Mount and automount units may now be created transiently
3235 (i.e. dynamically at runtime via the bus API, instead of requiring
3236 unit files in the file system).
3238 * systemd-mount is a new tool which may mount file systems – much like
3239 mount(8), optionally pulling in additional dependencies through
3240 transient .mount and .automount units. For example, this tool
3241 automatically runs fsck on a backing block device before mounting,
3242 and allows the automount logic to be used dynamically from the
3243 command line for establishing mount points. This tool is particularly
3244 useful when dealing with removable media, as it will ensure fsck is
3245 run – if necessary – before the first access and that the file system
3246 is quickly unmounted after each access by utilizing the automount
3247 logic. This maximizes the chance that the file system on the
3248 removable media stays in a clean state, and if it isn't in a clean
3249 state is fixed automatically.
3251 * LazyUnmount=yes option for mount units has been added to expose the
3252 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
3255 * /efi will be used as the mount point of the EFI boot partition, if
3256 the directory is present, and the mount point was not configured
3257 through other means (e.g. fstab). If /efi directory does not exist,
3258 /boot will be used as before. This makes it easier to automatically
3259 mount the EFI partition on systems where /boot is used for something
3262 * When operating on GPT disk images for containers, systemd-nspawn will
3263 now mount the ESP to /boot or /efi according to the same rules as PID
3264 1 running on a host. This allows tools like "bootctl" to operate
3265 correctly within such containers, in order to make container images
3266 bootable on physical systems.
3268 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
3270 * Two new user session targets have been added to support running
3271 graphical sessions under the systemd --user instance:
3272 graphical-session.target and graphical-session-pre.target. See
3273 systemd.special(7) for a description of how those targets should be
3276 * The vconsole initialization code has been significantly reworked to
3277 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
3278 support unicode keymaps. Font and keymap configuration will now be
3279 copied to all allocated virtual consoles.
3281 * FreeBSD's bhyve virtualization is now detected.
3283 * Information recorded in the journal for core dumps now includes the
3284 contents of /proc/mountinfo and the command line of the process at
3285 the top of the process hierarchy (which is usually the init process
3288 * systemd-journal-gatewayd learned the --directory= option to serve
3289 files from the specified location.
3291 * journalctl --root=… can be used to peruse the journal in the
3292 /var/log/ directories inside of a container tree. This is similar to
3293 the existing --machine= option, but does not require the container to
3296 * The hardware database has been extended to support
3297 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
3300 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
3301 specify the click rate for mice which include a horizontal wheel with
3302 a click rate that is different than the one for the vertical wheel.
3304 * systemd-run gained a new --wait option that makes service execution
3305 synchronous. (Specifically, the command will not return until the
3306 specified service binary exited.)
3308 * systemctl gained a new --wait option that causes the start command to
3309 wait until the units being started have terminated again.
3311 * A new journal output mode "short-full" has been added which displays
3312 timestamps with abbreviated English day names and adds a timezone
3313 suffix. Those timestamps include more information than the default
3314 "short" output mode, and can be passed directly to journalctl's
3315 --since= and --until= options.
3317 * /etc/resolv.conf will be bind-mounted into containers started by
3318 systemd-nspawn, if possible, so any changes to resolv.conf contents
3319 are automatically propagated to the container.
3321 * The number of instances for socket-activated services originating
3322 from a single IP address can be limited with
3323 MaxConnectionsPerSource=, extending the existing setting of
3326 * systemd-networkd gained support for vcan ("Virtual CAN") interface
3329 * .netdev and .network configuration can now be extended through
3332 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
3333 Segmentation Offload, Generic Receive Offload, Large Receive Offload
3334 can be enabled and disabled using the new UDPSegmentationOffload=,
3335 TCPSegmentationOffload=, GenericSegmentationOffload=,
3336 GenericReceiveOffload=, LargeReceiveOffload= options in the
3337 [Link] section of .link files.
3339 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
3340 Port VLAN ID can be configured for bridge devices using the new STP=,
3341 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
3342 section of .netdev files.
3344 * The route table to which routes received over DHCP or RA should be
3345 added can be configured with the new RouteTable= option in the [DHCP]
3346 and [IPv6AcceptRA] sections of .network files.
3348 * The Address Resolution Protocol can be disabled on links managed by
3349 systemd-networkd using the ARP=no setting in the [Link] section of
3352 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
3353 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
3354 encode information about the result and exit codes of the current
3355 service runtime cycle.
3357 * systemd-sysctl will now configure kernel parameters in the order
3358 they occur in the configuration files. This matches what sysctl
3359 has been traditionally doing.
3361 * kernel-install "plugins" that are executed to perform various
3362 tasks after a new kernel is added and before an old one is removed
3363 can now return a special value to terminate the procedure and
3364 prevent any later plugins from running.
3366 * Journald's SplitMode=login setting has been deprecated. It has been
3367 removed from documentation, and its use is discouraged. In a future
3368 release it will be completely removed, and made equivalent to current
3369 default of SplitMode=uid.
3371 * Storage=both option setting in /etc/systemd/coredump.conf has been
3372 removed. With fast LZ4 compression storing the core dump twice is not
3375 * The --share-system systemd-nspawn option has been replaced with an
3376 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
3377 this functionality is discouraged. In addition the variables
3378 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
3379 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
3380 individual namespaces.
3382 * "machinectl list" now shows the IP address of running containers in
3383 the output, as well as OS release information.
3385 * "loginctl list" now shows the TTY of each session in the output.
3387 * sd-bus gained new API calls sd_bus_track_set_recursive(),
3388 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
3389 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
3390 tracking objects in a "recursive" mode, where a single client can be
3391 counted multiple times, if it takes multiple references.
3393 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
3394 sd_bus_get_exit_on_disconnect(). They may be used to make a
3395 process using sd-bus automatically exit if the bus connection is
3398 * Bus clients of the service manager may now "pin" loaded units into
3399 memory, by taking an explicit reference on them. This is useful to
3400 ensure the client can retrieve runtime data about the service even
3401 after the service completed execution. Taking such a reference is
3402 available only for privileged clients and should be helpful to watch
3403 running services in a race-free manner, and in particular collect
3404 information about exit statuses and results.
3406 * The nss-resolve module has been changed to strictly return UNAVAIL
3407 when communication via D-Bus with resolved failed, and NOTFOUND when
3408 a lookup completed but was negative. This means it is now possible to
3409 neatly configure fallbacks using nsswitch.conf result checking
3410 expressions. Taking benefit of this, the new recommended
3411 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
3413 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
3415 * A new setting CtrlAltDelBurstAction= has been added to
3416 /etc/systemd/system.conf which may be used to configure the precise
3417 behaviour if the user on the console presses Ctrl-Alt-Del more often
3418 than 7 times in 2s. Previously this would unconditionally result in
3419 an expedited, immediate reboot. With this new setting the precise
3420 operation may be configured in more detail, and also turned off
3423 * In .netdev files two new settings RemoteChecksumTx= and
3424 RemoteChecksumRx= are now understood that permit configuring the
3425 remote checksumming logic for VXLAN networks.
3427 * The service manager learnt a new "invocation ID" concept for invoked
3428 services. Each runtime cycle of a service will get a new invocation
3429 ID (a 128bit random UUID) assigned that identifies the current
3430 run of the service uniquely and globally. A new invocation ID
3431 is generated each time a service starts up. The journal will store
3432 the invocation ID of a service along with any logged messages, thus
3433 making the invocation ID useful for matching the online runtime of a
3434 service with the offline log data it generated in a safe way without
3435 relying on synchronized timestamps. In many ways this new service
3436 invocation ID concept is similar to the kernel's boot ID concept that
3437 uniquely and globally identifies the runtime of each boot. The
3438 invocation ID of a service is passed to the service itself via an
3439 environment variable ($INVOCATION_ID). A new bus call
3440 GetUnitByInvocationID() has been added that is similar to GetUnit()
3441 but instead of retrieving the bus path for a unit by its name
3442 retrieves it by its invocation ID. The returned path is valid only as
3443 long as the passed invocation ID is current.
3445 * systemd-resolved gained a new "DNSStubListener" setting in
3446 resolved.conf. It either takes a boolean value or the special values
3447 "udp" and "tcp", and configures whether to enable the stub DNS
3448 listener on 127.0.0.53:53.
3450 * IP addresses configured via networkd may now carry additional
3451 configuration settings supported by the kernel. New options include:
3452 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
3453 PrefixRoute=, AutoJoin=.
3455 * The PAM configuration fragment file for "user@.service" shipped with
3456 systemd (i.e. the --user instance of systemd) has been stripped to
3457 the minimum necessary to make the system boot. Previously, it
3458 contained Fedora-specific stanzas that did not apply to other
3459 distributions. It is expected that downstream distributions add
3460 additional configuration lines, matching their needs to this file,
3461 using it only as rough template of what systemd itself needs. Note
3462 that this reduced fragment does not even include an invocation of
3463 pam_limits which most distributions probably want to add, even though
3464 systemd itself does not need it. (There's also the new build time
3465 option --with-pamconfdir=no to disable installation of the PAM
3468 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
3469 capability is now also dropped from its set (in addition to
3470 CAP_SYS_MKNOD as before).
3472 * In service unit files it is now possible to connect a specific named
3473 file descriptor with stdin/stdout/stdout of an executed service. The
3474 name may be specified in matching .socket units using the
3475 FileDescriptorName= setting.
3477 * A number of journal settings may now be configured on the kernel
3478 command line. Specifically, the following options are now understood:
3479 systemd.journald.max_level_console=,
3480 systemd.journald.max_level_store=,
3481 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
3482 systemd.journald.max_level_wall=.
3484 * "systemctl is-enabled --full" will now show by which symlinks a unit
3485 file is enabled in the unit dependency tree.
3487 * Support for VeraCrypt encrypted partitions has been added to the
3488 "cryptsetup" logic and /etc/crypttab.
3490 * systemd-detect-virt gained support for a new --private-users switch
3491 that checks whether the invoking processes are running inside a user
3492 namespace. Similar, a new special value "private-users" for the
3493 existing ConditionVirtualization= setting has been added, permitting
3494 skipping of specific units in user namespace environments.
3496 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
3497 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
3498 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
3499 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
3500 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
3501 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
3502 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
3503 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
3504 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
3505 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
3506 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
3507 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
3508 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
3509 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
3510 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
3511 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
3512 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
3513 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
3514 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
3515 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
3516 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
3517 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
3518 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
3519 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
3520 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
3521 Jędrzejewski-Szmek, Zeal Jagannatha
3523 — Santa Fe, 2016-11-03
3527 * In service units the various ExecXYZ= settings have been extended
3528 with an additional special character as first argument of the
3529 assigned value: if the character '+' is used the specified command
3530 line it will be run with full privileges, regardless of User=,
3531 Group=, CapabilityBoundingSet= and similar options. The effect is
3532 similar to the existing PermissionsStartOnly= option, but allows
3533 configuration of this concept for each executed command line
3536 * Services may now alter the service watchdog timeout at runtime by
3537 sending a WATCHDOG_USEC= message via sd_notify().
3539 * MemoryLimit= and related unit settings now optionally take percentage
3540 specifications. The percentage is taken relative to the amount of
3541 physical memory in the system (or in case of containers, the assigned
3542 amount of memory). This allows scaling service resources neatly with
3543 the amount of RAM available on the system. Similarly, systemd-logind's
3544 RuntimeDirectorySize= option now also optionally takes percentage
3547 * In similar fashion TasksMax= takes percentage values now, too. The
3548 value is taken relative to the configured maximum number of processes
3549 on the system. The per-service task maximum has been changed to 15%
3550 using this functionality. (Effectively this is an increase of 512 →
3551 4915 for service units, given the kernel's default pid_max setting.)
3553 * Calendar time specifications in .timer units now understand a ".."
3554 syntax for time ranges. Example: "4..7:10" may now be used for
3555 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
3558 * The InaccessableDirectories=, ReadOnlyDirectories= and
3559 ReadWriteDirectories= unit file settings have been renamed to
3560 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
3561 applied to all kinds of file nodes, and not just directories, with
3562 the exception of symlinks. Specifically these settings may now be
3563 used on block and character device nodes, UNIX sockets and FIFOS as
3564 well as regular files. The old names of these settings remain
3565 available for compatibility.
3567 * systemd will now log about all service processes it kills forcibly
3568 (using SIGKILL) because they remained after the clean shutdown phase
3569 of the service completed. This should help identifying services that
3570 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
3571 systemd-logind's configuration a similar log message is generated for
3572 processes killed at the end of each session due to this setting.
3574 * systemd will now set the $JOURNAL_STREAM environment variable for all
3575 services whose stdout/stderr are connected to the Journal (which
3576 effectively means by default: all services). The variable contains
3577 the device and inode number of the file descriptor used for
3578 stdout/stderr. This may be used by invoked programs to detect whether
3579 their stdout/stderr is connected to the Journal, in which case they
3580 can switch over to direct Journal communication, thus being able to
3581 pass extended, structured metadata along with their log messages. As
3582 one example, this is now used by glib's logging primitives.
3584 * When using systemd's default tmp.mount unit for /tmp, the mount point
3585 will now be established with the "nosuid" and "nodev" options. This
3586 avoids privilege escalation attacks that put traps and exploits into
3587 /tmp. However, this might cause problems if you e. g. put container
3588 images or overlays into /tmp; if you need this, override tmp.mount's
3589 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
3592 * systemd now supports the "memory" cgroup controller also on
3595 * The systemd-cgtop tool now optionally takes a control group path as
3596 command line argument. If specified, the control group list shown is
3597 limited to subgroups of that group.
3599 * The SystemCallFilter= unit file setting gained support for
3600 pre-defined, named system call filter sets. For example
3601 SystemCallFilter=@clock is now an effective way to make all clock
3602 changing-related system calls unavailable to a service. A number of
3603 similar pre-defined groups are defined. Writing system call filters
3604 for system services is simplified substantially with this new
3605 concept. Accordingly, all of systemd's own, long-running services now
3606 enable system call filtering based on this, by default.
3608 * A new service setting MemoryDenyWriteExecute= has been added, taking
3609 a boolean value. If turned on, a service may no longer create memory
3610 mappings that are writable and executable at the same time. This
3611 enhances security for services where this is enabled as it becomes
3612 harder to dynamically write and then execute memory in exploited
3613 service processes. This option has been enabled for all of systemd's
3614 own long-running services.
3616 * A new RestrictRealtime= service setting has been added, taking a
3617 boolean argument. If set the service's processes may no longer
3618 acquire realtime scheduling. This improves security as realtime
3619 scheduling may otherwise be used to easily freeze the system.
3621 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
3622 value. This may be used for requesting that the system manager inside
3623 of the container reports start-up completion to nspawn which then
3624 propagates this notification further to the service manager
3625 supervising nspawn itself. A related option NotifyReady= in .nspawn
3626 files has been added too. This functionality allows ordering of the
3627 start-up of multiple containers using the usual systemd ordering
3630 * machinectl gained a new command "stop" that is an alias for
3633 * systemd-resolved gained support for contacting DNS servers on
3634 link-local IPv6 addresses.
3636 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
3637 its caches. A method call for requesting the same operation has been
3638 added to the bus API too, and is made available via "systemd-resolve
3641 * systemd-resolve gained a new --status switch. If passed a brief
3642 summary of the used DNS configuration with per-interface information
3645 * resolved.conf gained a new Cache= boolean option, defaulting to
3646 on. If turned off local DNS caching is disabled. This comes with a
3647 performance penalty in particular when DNSSEC is enabled. Note that
3648 resolved disables its internal caching implicitly anyway, when the
3649 configured DNS server is on a host-local IP address such as ::1 or
3650 127.0.0.1, thus automatically avoiding double local caching.
3652 * systemd-resolved now listens on the local IP address 127.0.0.53:53
3653 for DNS requests. This improves compatibility with local programs
3654 that do not use the libc NSS or systemd-resolved's bus APIs for name
3655 resolution. This minimal DNS service is only available to local
3656 programs and does not implement the full DNS protocol, but enough to
3657 cover local DNS clients. A new, static resolv.conf file, listing just
3658 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
3659 now recommended to make /etc/resolv.conf a symlink to this file in
3660 order to route all DNS lookups to systemd-resolved, regardless if
3661 done via NSS, the bus API or raw DNS packets. Note that this local
3662 DNS service is not as fully featured as the libc NSS or
3663 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
3664 used to deliver link-local address information (as this implies
3665 sending a local interface index along), LLMNR/mDNS support via this
3666 interface is severely restricted. It is thus strongly recommended for
3667 all applications to use the libc NSS API or native systemd-resolved
3670 * systemd-networkd's bridge support learned a new setting
3671 VLANFiltering= for controlling VLAN filtering. Moreover a new section
3672 in .network files has been added for configuring VLAN bridging in
3673 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
3675 * systemd-networkd's IPv6 Router Advertisement code now makes use of
3676 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
3677 now be acquired without relying on DHCPv6. Two new options
3678 UseDomains= and UseDNS= have been added to configure this behaviour.
3680 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
3681 renamed IPv6AcceptRA=, without altering its behaviour. The old
3682 setting name remains available for compatibility reasons.
3684 * The systemd-networkd VTI/VTI6 tunneling support gained new options
3685 Key=, InputKey= and OutputKey=.
3687 * systemd-networkd gained support for VRF ("Virtual Routing Function")
3688 interface configuration.
3690 * "systemctl edit" may now be used to create new unit files by
3691 specifying the --force switch.
3693 * sd-event gained a new function sd_event_get_iteration() for
3694 requesting the current iteration counter of the event loop. It starts
3695 at zero and is increased by one with each event loop iteration.
3697 * A new rpm macro %systemd_ordering is provided by the macros.systemd
3698 file. It can be used in lieu of %systemd_requires in packages which
3699 don't use any systemd functionality and are intended to be installed
3700 in minimal containers without systemd present. This macro provides
3701 ordering dependencies to ensure that if the package is installed in
3702 the same rpm transaction as systemd, systemd will be installed before
3703 the scriptlets for the package are executed, allowing unit presets
3706 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
3707 been added to simplify packaging of generators.
3709 * The os-release file gained VERSION_CODENAME field for the
3710 distribution nickname (e.g. VERSION_CODENAME=woody).
3712 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
3713 can be set to disable parsing of metadata and the creation
3714 of persistent symlinks for that device.
3716 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
3717 to make them available to logged-in users has been reverted.
3719 * Much of the common code of the various systemd components is now
3720 built into an internal shared library libsystemd-shared-231.so
3721 (incorporating the systemd version number in the name, to be updated
3722 with future releases) that the components link to. This should
3723 decrease systemd footprint both in memory during runtime and on
3724 disk. Note that the shared library is not for public use, and is
3725 neither API nor ABI stable, but is likely to change with every new
3726 released update. Packagers need to make sure that binaries
3727 linking to libsystemd-shared.so are updated in step with the
3730 * Configuration for "mkosi" is now part of the systemd
3731 repository. mkosi is a tool to easily build legacy-free OS images,
3732 and is available on github: https://github.com/systemd/mkosi. If
3733 "mkosi" is invoked in the build tree a new raw OS image is generated
3734 incorporating the systemd sources currently being worked on and a
3735 clean, fresh distribution installation. The generated OS image may be
3736 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
3737 UEFI PC. This functionality is particularly useful to easily test
3738 local changes made to systemd in a pristine, defined environment. See
3739 doc/HACKING for details.
3741 * configure learned the --with-support-url= option to specify the
3742 distribution's bugtracker.
3744 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
3745 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
3746 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
3747 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
3748 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
3749 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
3750 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
3751 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
3752 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
3753 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
3754 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
3755 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
3756 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
3757 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
3758 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
3759 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
3760 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
3761 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
3762 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
3764 — Berlin, 2016-07-25
3768 * DNSSEC is now turned on by default in systemd-resolved (in
3769 "allow-downgrade" mode), but may be turned off during compile time by
3770 passing "--with-default-dnssec=no" to "configure" (and of course,
3771 during runtime with DNSSEC= in resolved.conf). We recommend
3772 downstreams to leave this on at least during development cycles and
3773 report any issues with the DNSSEC logic upstream. We are very
3774 interested in collecting feedback about the DNSSEC validator and its
3775 limitations in the wild. Note however, that DNSSEC support is
3776 probably nothing downstreams should turn on in stable distros just
3777 yet, as it might create incompatibilities with a few DNS servers and
3778 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
3779 automatically whenever we detect such incompatible setups, but there
3780 might be systems we do not cover yet. Hence: please help us testing
3781 the DNSSEC code, leave this on where you can, report back, but then
3782 again don't consider turning this on in your stable, LTS or
3783 production release just yet. (Note that you have to enable
3784 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
3785 and its DNSSEC mode for host name resolution from local
3788 * systemd-resolve conveniently resolves DANE records with the --tlsa
3789 option and OPENPGPKEY records with the --openpgp option. It also
3790 supports dumping raw DNS record data via the new --raw= switch.
3792 * systemd-logind will now by default terminate user processes that are
3793 part of the user session scope unit (session-XX.scope) when the user
3794 logs out. This behavior is controlled by the KillUserProcesses=
3795 setting in logind.conf, and the previous default of "no" is now
3796 changed to "yes". This means that user sessions will be properly
3797 cleaned up after, but additional steps are necessary to allow
3798 intentionally long-running processes to survive logout.
3800 While the user is logged in at least once, user@.service is running,
3801 and any service that should survive the end of any individual login
3802 session can be started at a user service or scope using systemd-run.
3803 systemd-run(1) man page has been extended with an example which shows
3804 how to run screen in a scope unit underneath user@.service. The same
3805 command works for tmux.
3807 After the user logs out of all sessions, user@.service will be
3808 terminated too, by default, unless the user has "lingering" enabled.
3809 To effectively allow users to run long-term tasks even if they are
3810 logged out, lingering must be enabled for them. See loginctl(1) for
3811 details. The default polkit policy was modified to allow users to
3812 set lingering for themselves without authentication.
3814 Previous defaults can be restored at compile time by the
3815 --without-kill-user-processes option to "configure".
3817 * systemd-logind gained new configuration settings SessionsMax= and
3818 InhibitorsMax=, both with a default of 8192. It will not register new
3819 user sessions or inhibitors above this limit.
3821 * systemd-logind will now reload configuration on SIGHUP.
3823 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
3824 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
3825 enable. Also, support for the "io" cgroup controller in the unified
3826 hierarchy has been added, so that the "memory", "pids" and "io" are
3827 now the controllers that are supported on the unified hierarchy.
3829 WARNING: it is not possible to use previous systemd versions with
3830 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
3831 is necessary to also update systemd in the initramfs if using the
3832 unified hierarchy. An updated SELinux policy is also required.
3834 * LLDP support has been extended, and both passive (receive-only) and
3835 active (sender) modes are supported. Passive mode ("routers-only") is
3836 enabled by default in systemd-networkd. Active LLDP mode is enabled
3837 by default for containers on the internal network. The "networkctl
3838 lldp" command may be used to list information gathered. "networkctl
3839 status" will also show basic LLDP information on connected peers now.
3841 * The IAID and DUID unique identifier sent in DHCP requests may now be
3842 configured for the system and each .network file managed by
3843 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
3845 * systemd-networkd gained support for configuring proxy ARP support for
3846 each interface, via the ProxyArp= setting in .network files. It also
3847 gained support for configuring the multicast querier feature of
3848 bridge devices, via the new MulticastQuerier= setting in .netdev
3849 files. Similarly, snooping on the IGMP traffic can be controlled
3850 via the new setting MulticastSnooping=.
3852 A new setting PreferredLifetime= has been added for addresses
3853 configured in .network file to configure the lifetime intended for an
3856 The systemd-networkd DHCP server gained the option EmitRouter=, which
3857 defaults to yes, to configure whether the DHCP Option 3 (Router)
3860 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
3861 systemd-socket-activate and installed into /usr/bin. It is now fully
3864 * systemd-journald now uses separate threads to flush changes to disk
3865 when closing journal files, thus reducing impact of slow disk I/O on
3866 logging performance.
3868 * The sd-journal API gained two new calls
3869 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
3870 can be used to open journal files using file descriptors instead of
3871 file or directory paths. sd_journal_open_container() has been
3872 deprecated, sd_journal_open_directory_fd() should be used instead
3873 with the flag SD_JOURNAL_OS_ROOT.
3875 * journalctl learned a new output mode "-o short-unix" that outputs log
3876 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
3877 UTC). It also gained support for a new --no-hostname setting to
3878 suppress the hostname column in the family of "short" output modes.
3880 * systemd-ask-password now optionally skips printing of the password to
3881 stdout with --no-output which can be useful in scripts.
3883 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
3884 (devices tagged with ID_MAKER_TOOL) are now tagged with
3885 "uaccess" and are available to logged in users.
3887 * The DeviceAllow= unit setting now supports specifiers (with "%").
3889 * "systemctl show" gained a new --value switch, which allows print a
3890 only the contents of a specific unit property, without also printing
3891 the property's name. Similar support was added to "show*" verbs
3892 of loginctl and machinectl that output "key=value" lists.
3894 * A new unit type "generated" was added for files dynamically generated
3895 by generator tools. Similarly, a new unit type "transient" is used
3896 for unit files created using the runtime API. "systemctl enable" will
3897 refuse to operate on such files.
3899 * A new command "systemctl revert" has been added that may be used to
3900 revert to the vendor version of a unit file, in case local changes
3901 have been made by adding drop-ins or overriding the unit file.
3903 * "machinectl clean" gained a new verb to automatically remove all or
3904 just hidden container images.
3906 * systemd-tmpfiles gained support for a new line type "e" for emptying
3907 directories, if they exist, without creating them if they don't.
3909 * systemd-nspawn gained support for automatically patching the UID/GIDs
3910 of the owners and the ACLs of all files and directories in a
3911 container tree to match the UID/GID user namespacing range selected
3912 for the container invocation. This mode is enabled via the new
3913 --private-users-chown switch. It also gained support for
3914 automatically choosing a free, previously unused UID/GID range when
3915 starting a container, via the new --private-users=pick setting (which
3916 implies --private-users-chown). Together, these options for the first
3917 time make user namespacing for nspawn containers fully automatic and
3918 thus deployable. The systemd-nspawn@.service template unit file has
3919 been changed to use this functionality by default.
3921 * systemd-nspawn gained a new --network-zone= switch, that allows
3922 creating ad-hoc virtual Ethernet links between multiple containers,
3923 that only exist as long as at least one container referencing them is
3924 running. This allows easy connecting of multiple containers with a
3925 common link that implements an Ethernet broadcast domain. Each of
3926 these network "zones" may be named relatively freely by the user, and
3927 may be referenced by any number of containers, but each container may
3928 only reference one of these "zones". On the lower level, this is
3929 implemented by an automatically managed bridge network interface for
3930 each zone, that is created when the first container referencing its
3931 zone is created and removed when the last one referencing its zone
3934 * The default start timeout may now be configured on the kernel command
3935 line via systemd.default_timeout_start_sec=. It was already
3936 configurable via the DefaultTimeoutStartSec= option in
3937 /etc/systemd/system.conf.
3939 * Socket units gained a new TriggerLimitIntervalSec= and
3940 TriggerLimitBurst= setting to configure a limit on the activation
3941 rate of the socket unit.
3943 * The LimitNICE= setting now optionally takes normal UNIX nice values
3944 in addition to the raw integer limit value. If the specified
3945 parameter is prefixed with "+" or "-" and is in the range -20..19 the
3946 value is understood as UNIX nice value. If not prefixed like this it
3947 is understood as raw RLIMIT_NICE limit.
3949 * Note that the effect of the PrivateDevices= unit file setting changed
3950 slightly with this release: the per-device /dev file system will be
3951 mounted read-only from this version on, and will have "noexec"
3952 set. This (minor) change of behavior might cause some (exceptional)
3953 legacy software to break, when PrivateDevices=yes is set for its
3954 service. Please leave PrivateDevices= off if you run into problems
3957 * systemd-bootchart has been split out to a separate repository:
3958 https://github.com/systemd/systemd-bootchart
3960 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
3961 merged into the kernel in its current form.
3963 * The compatibility libraries libsystemd-daemon.so,
3964 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
3965 which have been deprecated since systemd-209 have been removed along
3966 with the corresponding pkg-config files. All symbols provided by
3967 those libraries are provided by libsystemd.so.
3969 * The Capabilities= unit file setting has been removed (it is ignored
3970 for backwards compatibility). AmbientCapabilities= and
3971 CapabilityBoundingSet= should be used instead.
3973 * A new special target has been added, initrd-root-device.target,
3974 which creates a synchronization point for dependencies of the root
3975 device in early userspace. Initramfs builders must ensure that this
3976 target is now included in early userspace.
3978 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
3979 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
3980 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
3981 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
3982 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
3983 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
3984 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
3985 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
3986 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
3987 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
3988 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
3989 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
3990 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
3991 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
3992 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
3993 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
3994 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
3995 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
3996 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
3997 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
3998 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
3999 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
4000 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
4001 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
4002 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
4005 — Fairfax, 2016-05-21
4009 * The systemd-resolved DNS resolver service has gained a substantial
4010 set of new features, most prominently it may now act as a DNSSEC
4011 validating stub resolver. DNSSEC mode is currently turned off by
4012 default, but is expected to be turned on by default in one of the
4013 next releases. For now, we invite everybody to test the DNSSEC logic
4014 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
4015 service also gained a full set of D-Bus interfaces, including calls
4016 to configure DNS and DNSSEC settings per link (for use by external
4017 network management software). systemd-resolved and systemd-networkd
4018 now distinguish between "search" and "routing" domains. The former
4019 are used to qualify single-label names, the latter are used purely
4020 for routing lookups within certain domains to specific links.
4021 resolved now also synthesizes RRs for all entries from /etc/hosts.
4023 * The systemd-resolve tool (which is a client utility for
4024 systemd-resolved) has been improved considerably and is now fully
4025 supported and documented. Hence it has moved from /usr/lib/systemd to
4028 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
4031 * The coredump collection logic has been reworked: when a coredump is
4032 collected it is now written to disk, compressed and processed
4033 (including stacktrace extraction) from a new instantiated service
4034 systemd-coredump@.service, instead of directly from the
4035 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
4036 processing large coredumps can take up a substantial amount of
4037 resources and time, and this previously happened entirely outside of
4038 systemd's service supervision. With the new logic the core_pattern
4039 hook only does minimal metadata collection before passing off control
4040 to the new instantiated service, which is configured with a time
4041 limit, a nice level and other settings to minimize negative impact on
4042 the rest of the system. Also note that the new logic will honour the
4043 RLIMIT_CORE setting of the crashed process, which now allows users
4044 and processes to turn off coredumping for their processes by setting
4047 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
4048 and all forked processes by default. Previously, PID 1 would leave
4049 the setting at "0" for all processes, as set by the kernel. Note that
4050 the resource limit traditionally has no effect on the generated
4051 coredumps on the system if the /proc/sys/kernel/core_pattern hook
4052 logic is used. Since the limit is now honoured (see above) its
4053 default has been changed so that the coredumping logic is enabled by
4054 default for all processes, while allowing specific opt-out.
4056 * When the stacktrace is extracted from processes of system users, this
4057 is now done as "systemd-coredump" user, in order to sandbox this
4058 potentially security sensitive parsing operation. (Note that when
4059 processing coredumps of normal users this is done under the user ID
4060 of process that crashed, as before.) Packagers should take notice
4061 that it is now necessary to create the "systemd-coredump" system user
4062 and group at package installation time.
4064 * The systemd-activate socket activation testing tool gained support
4065 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
4066 and --seqpacket switches. It also has been extended to support both
4067 new-style and inetd-style file descriptor passing. Use the new
4068 --inetd switch to request inetd-style file descriptor passing.
4070 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
4071 variable, which takes a boolean value. If set to false, ANSI color
4072 output is disabled in the tools even when run on a terminal that
4075 * The VXLAN support in networkd now supports two new settings
4076 DestinationPort= and PortRange=.
4078 * A new systemd.machine_id= kernel command line switch has been added,
4079 that may be used to set the machine ID in /etc/machine-id if it is
4080 not initialized yet. This command line option has no effect if the
4081 file is already initialized.
4083 * systemd-nspawn gained a new --as-pid2 switch that invokes any
4084 specified command line as PID 2 rather than PID 1 in the
4085 container. In this mode PID 1 is a minimal stub init process that
4086 implements the special POSIX and Linux semantics of PID 1 regarding
4087 signal and child process management. Note that this stub init process
4088 is implemented in nspawn itself and requires no support from the
4089 container image. This new logic is useful to support running
4090 arbitrary commands in the container, as normal processes are
4091 generally not prepared to run as PID 1.
4093 * systemd-nspawn gained a new --chdir= switch for setting the current
4094 working directory for the process started in the container.
4096 * "journalctl /dev/sda" will now output all kernel log messages for
4097 specified device from the current boot, in addition to all devices
4098 that are parents of it. This should make log output about devices
4099 pretty useful, as long as kernel drivers attach enough metadata to
4100 the log messages. (The usual SATA drivers do.)
4102 * The sd-journal API gained two new calls
4103 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
4104 that report whether log data from /run or /var has been found.
4106 * journalctl gained a new switch "--fields" that prints all journal
4107 record field names currently in use in the journal. This is backed
4108 by two new sd-journal API calls sd_journal_enumerate_fields() and
4109 sd_journal_restart_fields().
4111 * Most configurable timeouts in systemd now expect an argument of
4112 "infinity" to turn them off, instead of "0" as before. The semantics
4113 from now on is that a timeout of "0" means "now", and "infinity"
4114 means "never". To maintain backwards compatibility, "0" continues to
4115 turn off previously existing timeout settings.
4117 * "systemctl reload-or-try-restart" has been renamed to "systemctl
4118 try-reload-or-restart" to clarify what it actually does: the "try"
4119 logic applies to both reloading and restarting, not just restarting.
4120 The old name continues to be accepted for compatibility.
4122 * On boot-up, when PID 1 detects that the system clock is behind the
4123 release date of the systemd version in use, the clock is now set
4124 to the latter. Previously, this was already done in timesyncd, in order
4125 to avoid running with clocks set to the various clock epochs such as
4126 1902, 1938 or 1970. With this change the logic is now done in PID 1
4127 in addition to timesyncd during early boot-up, so that it is enforced
4128 before the first process is spawned by systemd. Note that the logic
4129 in timesyncd remains, as it is more comprehensive and ensures
4130 clock monotonicity by maintaining a persistent timestamp file in
4131 /var. Since /var is generally not available in earliest boot or the
4132 initrd, this part of the logic remains in timesyncd, and is not done
4135 * Support for tweaking details in net_cls.class_id through the
4136 NetClass= configuration directive has been removed, as the kernel
4137 people have decided to deprecate that controller in cgroup v2.
4138 Userspace tools such as nftables are moving over to setting rules
4139 that are specific to the full cgroup path of a task, which obsoletes
4140 these controllers anyway. The NetClass= directive is kept around for
4141 legacy compatibility reasons. For a more in-depth description of the
4142 kernel change, please refer to the respective upstream commit:
4144 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
4146 * A new service setting RuntimeMaxSec= has been added that may be used
4147 to specify a maximum runtime for a service. If the timeout is hit, the
4148 service is terminated and put into a failure state.
4150 * A new service setting AmbientCapabilities= has been added. It allows
4151 configuration of additional Linux process capabilities that are
4152 passed to the activated processes. This is only available on very
4155 * The process resource limit settings in service units may now be used
4156 to configure hard and soft limits individually.
4158 * The various libsystemd APIs such as sd-bus or sd-event now publicly
4159 expose support for gcc's __attribute__((cleanup())) C extension.
4160 Specifically, for many object destructor functions alternative
4161 versions have been added that have names suffixed with "p" and take a
4162 pointer to a pointer to the object to destroy, instead of just a
4163 pointer to the object itself. This is useful because these destructor
4164 functions may be used directly as parameters to the cleanup
4165 construct. Internally, systemd has been a heavy user of this GCC
4166 extension for a long time, and with this change similar support is
4167 now available to consumers of the library outside of systemd. Note
4168 that by using this extension in your sources compatibility with old
4169 and strictly ANSI compatible C compilers is lost. However, all gcc or
4170 LLVM versions of recent years support this extension.
4172 * Timer units gained support for a new setting RandomizedDelaySec= that
4173 allows configuring some additional randomized delay to the configured
4174 time. This is useful to spread out timer events to avoid load peaks in
4175 clusters or larger setups.
4177 * Calendar time specifications now support sub-second accuracy.
4179 * Socket units now support listening on SCTP and UDP-lite protocol
4182 * The sd-event API now comes with a full set of man pages.
4184 * Older versions of systemd contained experimental support for
4185 compressing journal files and coredumps with the LZ4 compressor that
4186 was not compatible with the lz4 binary (due to API limitations of the
4187 lz4 library). This support has been removed; only support for files
4188 compatible with the lz4 binary remains. This LZ4 logic is now
4189 officially supported and no longer considered experimental.
4191 * The dkr image import logic has been removed again from importd. dkr's
4192 micro-services focus doesn't fit into the machine image focus of
4193 importd, and quickly got out of date with the upstream dkr API.
4195 * Creation of the /run/lock/lockdev/ directory was dropped from
4196 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
4197 been available for many years. If you still need this, you need to
4198 create your own tmpfiles.d config file with:
4200 d /run/lock/lockdev 0775 root lock -
4202 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
4203 and RebootArgument= have been moved from the [Service] section of
4204 unit files to [Unit], and they are now supported on all unit types,
4205 not just service units. Of course, systemd will continue to
4206 understand these settings also at the old location, in order to
4207 maintain compatibility.
4209 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
4210 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
4211 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
4212 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
4213 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
4214 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
4215 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
4216 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
4217 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
4218 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
4219 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
4220 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
4221 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
4222 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
4223 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
4224 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
4225 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
4226 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
4227 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
4229 — Berlin, 2016-02-11
4233 * A number of properties previously only settable in unit
4234 files are now also available as properties to set when
4235 creating transient units programmatically via the bus, as it
4236 is exposed with systemd-run's --property=
4237 setting. Specifically, these are: SyslogIdentifier=,
4238 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
4239 EnvironmentFile=, ReadWriteDirectories=,
4240 ReadOnlyDirectories=, InaccessibleDirectories=,
4241 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
4243 * When creating transient services via the bus API it is now
4244 possible to pass in a set of file descriptors to use as
4245 STDIN/STDOUT/STDERR for the invoked process.
4247 * Slice units may now be created transiently via the bus APIs,
4248 similar to the way service and scope units may already be
4249 created transiently.
4251 * Wherever systemd expects a calendar timestamp specification
4252 (like in journalctl's --since= and --until= switches) UTC
4253 timestamps are now supported. Timestamps suffixed with "UTC"
4254 are now considered to be in Universal Time Coordinated
4255 instead of the local timezone. Also, timestamps may now
4256 optionally be specified with sub-second accuracy. Both of
4257 these additions also apply to recurring calendar event
4258 specification, such as OnCalendar= in timer units.
4260 * journalctl gained a new "--sync" switch that asks the
4261 journal daemon to write all so far unwritten log messages to
4262 disk and sync the files, before returning.
4264 * systemd-tmpfiles learned two new line types "q" and "Q" that
4265 operate like "v", but also set up a basic btrfs quota
4266 hierarchy when used on a btrfs file system with quota
4269 * tmpfiles' "v", "q" and "Q" will now create a plain directory
4270 instead of a subvolume (even on a btrfs file system) if the
4271 root directory is a plain directory, and not a
4272 subvolume. This should simplify things with certain chroot()
4273 environments which are not aware of the concept of btrfs
4276 * systemd-detect-virt gained a new --chroot switch to detect
4277 whether execution takes place in a chroot() environment.
4279 * CPUAffinity= now takes CPU index ranges in addition to
4282 * The various memory-related resource limit settings (such as
4283 LimitAS=) now understand the usual K, M, G, ... suffixes to
4284 the base of 1024 (IEC). Similar, the time-related resource
4285 limit settings understand the usual min, h, day, ...
4288 * There's a new system.conf setting DefaultTasksMax= to
4289 control the default TasksMax= setting for services and
4290 scopes running on the system. (TasksMax= is the primary
4291 setting that exposes the "pids" cgroup controller on systemd
4292 and was introduced in the previous systemd release.) The
4293 setting now defaults to 512, which means services that are
4294 not explicitly configured otherwise will only be able to
4295 create 512 processes or threads at maximum, from this
4296 version on. Note that this means that thread- or
4297 process-heavy services might need to be reconfigured to set
4298 TasksMax= to a higher value. It is sufficient to set
4299 TasksMax= in these specific unit files to a higher value, or
4300 even "infinity". Similar, there's now a logind.conf setting
4301 UserTasksMax= that defaults to 4096 and limits the total
4302 number of processes or tasks each user may own
4303 concurrently. nspawn containers also have the TasksMax=
4304 value set by default now, to 8192. Note that all of this
4305 only has an effect if the "pids" cgroup controller is
4306 enabled in the kernel. The general benefit of these changes
4307 should be a more robust and safer system, that provides a
4308 certain amount of per-service fork() bomb protection.
4310 * systemd-nspawn gained the new --network-veth-extra= switch
4311 to define additional and arbitrarily-named virtual Ethernet
4312 links between the host and the container.
4314 * A new service execution setting PassEnvironment= has been
4315 added that allows importing select environment variables
4316 from PID1's environment block into the environment block of
4319 * Timer units gained support for a new RemainAfterElapse=
4320 setting which takes a boolean argument. It defaults to on,
4321 exposing behaviour unchanged to previous releases. If set to
4322 off, timer units are unloaded after they elapsed if they
4323 cannot elapse again. This is particularly useful for
4324 transient timer units, which shall not stay around longer
4325 than until they first elapse.
4327 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
4328 default now (the kernel default is 16). This is beneficial
4329 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
4330 allows substantially larger numbers of queued
4331 datagrams. This should increase the capability of systemd to
4332 parallelize boot-up, as logging and sd_notify() are unlikely
4333 to stall execution anymore. If you need to change the value
4334 from the new defaults, use the usual sysctl.d/ snippets.
4336 * The compression framing format used by the journal or
4337 coredump processing has changed to be in line with what the
4338 official LZ4 tools generate. LZ4 compression support in
4339 systemd was considered unsupported previously, as the format
4340 was not compatible with the normal tools. With this release
4341 this has changed now, and it is hence safe for downstream
4342 distributions to turn it on. While not compressing as well
4343 as the XZ, LZ4 is substantially faster, which makes
4344 it a good default choice for the compression logic in the
4345 journal and in coredump handling.
4347 * Any reference to /etc/mtab has been dropped from
4348 systemd. The file has been obsolete since a while, but
4349 systemd refused to work on systems where it was incorrectly
4350 set up (it should be a symlink or non-existent). Please make
4351 sure to update to util-linux 2.27.1 or newer in conjunction
4352 with this systemd release, which also drops any reference to
4353 /etc/mtab. If you maintain a distribution make sure that no
4354 software you package still references it, as this is a
4355 likely source of bugs. There's also a glibc bug pending,
4356 asking for removal of any reference to this obsolete file:
4358 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
4360 Note that only util-linux versions built with
4361 --enable-libmount-force-mountinfo are supported.
4363 * Support for the ".snapshot" unit type has been removed. This
4364 feature turned out to be little useful and little used, and
4365 has now been removed from the core and from systemctl.
4367 * The dependency types RequiresOverridable= and
4368 RequisiteOverridable= have been removed from systemd. They
4369 have been used only very sparingly to our knowledge and
4370 other options that provide a similar effect (such as
4371 systemctl --mode=ignore-dependencies) are much more useful
4372 and commonly used. Moreover, they were only half-way
4373 implemented as the option to control behaviour regarding
4374 these dependencies was never added to systemctl. By removing
4375 these dependency types the execution engine becomes a bit
4376 simpler. Unit files that use these dependencies should be
4377 changed to use the non-Overridable dependency types
4378 instead. In fact, when parsing unit files with these
4379 options, that's what systemd will automatically convert them
4380 too, but it will also warn, asking users to fix the unit
4381 files accordingly. Removal of these dependency types should
4382 only affect a negligible number of unit files in the wild.
4384 * Behaviour of networkd's IPForward= option changed
4385 (again). It will no longer maintain a per-interface setting,
4386 but propagate one way from interfaces where this is enabled
4387 to the global kernel setting. The global setting will be
4388 enabled when requested by a network that is set up, but
4389 never be disabled again. This change was made to make sure
4390 IPv4 and IPv6 behaviour regarding packet forwarding is
4391 similar (as the Linux IPv6 stack does not support
4392 per-interface control of this setting) and to minimize
4395 * In unit files the behaviour of %u, %U, %h, %s has
4396 changed. These specifiers will now unconditionally resolve
4397 to the various user database fields of the user that the
4398 systemd instance is running as, instead of the user
4399 configured in the specific unit via User=. Note that this
4400 effectively doesn't change much, as resolving of these
4401 specifiers was already turned off in the --system instance
4402 of systemd, as we cannot do NSS lookups from PID 1. In the
4403 --user instance of systemd these specifiers where correctly
4404 resolved, but hardly made any sense, since the user instance
4405 lacks privileges to do user switches anyway, and User= is
4406 hence useless. Moreover, even in the --user instance of
4407 systemd behaviour was awkward as it would only take settings
4408 from User= assignment placed before the specifier into
4409 account. In order to unify and simplify the logic around
4410 this the specifiers will now always resolve to the
4411 credentials of the user invoking the manager (which in case
4412 of PID 1 is the root user).
4414 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
4415 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
4416 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
4417 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
4418 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
4419 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
4420 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
4421 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
4422 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
4423 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
4424 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
4425 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
4426 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
4427 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
4430 — Berlin, 2015-11-18
4434 * systemd now depends on util-linux v2.27. More specifically,
4435 the newly added mount monitor feature in libmount now
4436 replaces systemd's former own implementation.
4438 * libmount mandates /etc/mtab not to be regular file, and
4439 systemd now enforces this condition at early boot.
4440 /etc/mtab has been deprecated and warned about for a very
4441 long time, so systems running systemd should already have
4442 stopped having this file around as anything else than a
4443 symlink to /proc/self/mounts.
4445 * Support for the "pids" cgroup controller has been added. It
4446 allows accounting the number of tasks in a cgroup and
4447 enforcing limits on it. This adds two new setting
4448 TasksAccounting= and TasksMax= to each unit, as well as a
4449 global option DefaultTasksAccounting=.
4451 * Support for the "net_cls" cgroup controller has been added.
4452 It allows assigning a net class ID to each task in the
4453 cgroup, which can then be used in firewall rules and traffic
4454 shaping configurations. Note that the kernel netfilter net
4455 class code does not currently work reliably for ingress
4456 packets on unestablished sockets.
4458 This adds a new config directive called NetClass= to CGroup
4459 enabled units. Allowed values are positive numbers for fixed
4460 assignments and "auto" for picking a free value
4463 * 'systemctl is-system-running' now returns 'offline' if the
4464 system is not booted with systemd. This command can now be
4465 used as a substitute for 'systemd-notify --booted'.
4467 * Watchdog timeouts have been increased to 3 minutes for all
4468 in-tree service files. Apparently, disk IO issues are more
4469 frequent than we hoped, and user reported >1 minute waiting
4472 * 'machine-id-commit' functionality has been merged into
4473 'machine-id-setup --commit'. The separate binary has been
4476 * The WorkingDirectory= directive in unit files may now be set
4477 to the special value '~'. In this case, the working
4478 directory is set to the home directory of the user
4479 configured in User=.
4481 * "machinectl shell" will now open the shell in the home
4482 directory of the selected user by default.
4484 * The CrashChVT= configuration file setting is renamed to
4485 CrashChangeVT=, following our usual logic of not
4486 abbreviating unnecessarily. The old directive is still
4487 supported for compat reasons. Also, this directive now takes
4488 an integer value between 1 and 63, or a boolean value. The
4489 formerly supported '-1' value for disabling stays around for
4492 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
4493 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
4494 RootDirectory= properties can now be set for transient
4497 * The systemd-analyze tool gained a new "set-log-target" verb
4498 to change the logging target the system manager logs to
4499 dynamically during runtime. This is similar to how
4500 "systemd-analyze set-log-level" already changes the log
4503 * In nspawn /sys is now mounted as tmpfs, with only a selected
4504 set of subdirectories mounted in from the real sysfs. This
4505 enhances security slightly, and is useful for ensuring user
4506 namespaces work correctly.
4508 * Support for USB FunctionFS activation has been added. This
4509 allows implementation of USB gadget services that are
4510 activated as soon as they are requested, so that they don't
4511 have to run continuously, similar to classic socket
4514 * The "systemctl exit" command now optionally takes an
4515 additional parameter that sets the exit code to return from
4516 the systemd manager when exiting. This is only relevant when
4517 running the systemd user instance, or when running the
4518 system instance in a container.
4520 * sd-bus gained the new API calls sd_bus_path_encode_many()
4521 and sd_bus_path_decode_many() that allow easy encoding and
4522 decoding of multiple identifier strings inside a D-Bus
4523 object path. Another new call sd_bus_default_flush_close()
4524 has been added to flush and close per-thread default
4527 * systemd-cgtop gained support for a -M/--machine= switch to
4528 show the control groups within a certain container only.
4530 * "systemctl kill" gained support for an optional --fail
4531 switch. If specified the requested operation will fail of no
4532 processes have been killed, because the unit had no
4533 processes attached, or similar.
4535 * A new systemd.crash_reboot=1 kernel command line option has
4536 been added that triggers a reboot after crashing. This can
4537 also be set through CrashReboot= in systemd.conf.
4539 * The RuntimeDirectory= setting now understands unit
4540 specifiers like %i or %f.
4542 * A new (still internal) library API sd-ipv4acd has been added,
4543 that implements address conflict detection for IPv4. It's
4544 based on code from sd-ipv4ll, and will be useful for
4545 detecting DHCP address conflicts.
4547 * File descriptors passed during socket activation may now be
4548 named. A new API sd_listen_fds_with_names() is added to
4549 access the names. The default names may be overridden,
4550 either in the .socket file using the FileDescriptorName=
4551 parameter, or by passing FDNAME= when storing the file
4552 descriptors using sd_notify().
4554 * systemd-networkd gained support for:
4556 - Setting the IPv6 Router Advertisement settings via
4557 IPv6AcceptRouterAdvertisements= in .network files.
4559 - Configuring the HelloTimeSec=, MaxAgeSec= and
4560 ForwardDelaySec= bridge parameters in .netdev files.
4562 - Configuring PreferredSource= for static routes in
4565 * The "ask-password" framework used to query for LUKS harddisk
4566 passwords or SSL passwords during boot gained support for
4567 caching passwords in the kernel keyring, if it is
4568 available. This makes sure that the user only has to type in
4569 a passphrase once if there are multiple objects to unlock
4570 with the same one. Previously, such password caching was
4571 available only when Plymouth was used; this moves the
4572 caching logic into the systemd codebase itself. The
4573 "systemd-ask-password" utility gained a new --keyname=
4574 switch to control which kernel keyring key to use for
4575 caching a password in. This functionality is also useful for
4576 enabling display managers such as gdm to automatically
4577 unlock the user's GNOME keyring if its passphrase, the
4578 user's password and the harddisk password are the same, if
4579 gdm-autologin is used.
4581 * When downloading tar or raw images using "machinectl
4582 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
4583 file is now also downloaded, if it is available and stored
4584 next to the image file.
4586 * Units of type ".socket" gained a new boolean setting
4587 Writable= which is only useful in conjunction with
4588 ListenSpecial=. If true, enables opening the specified
4589 special file in O_RDWR mode rather than O_RDONLY mode.
4591 * systemd-rfkill has been reworked to become a singleton
4592 service that is activated through /dev/rfkill on each rfkill
4593 state change and saves the settings to disk. This way,
4594 systemd-rfkill is now compatible with devices that exist
4595 only intermittendly, and even restores state if the previous
4596 system shutdown was abrupt rather than clean.
4598 * The journal daemon gained support for vacuuming old journal
4599 files controlled by the number of files that shall remain,
4600 in addition to the already existing control by size and by
4601 date. This is useful as journal interleaving performance
4602 degrades with too many separate journal files, and allows
4603 putting an effective limit on them. The new setting defaults
4604 to 100, but this may be changed by setting SystemMaxFiles=
4605 and RuntimeMaxFiles= in journald.conf. Also, the
4606 "journalctl" tool gained the new --vacuum-files= switch to
4607 manually vacuum journal files to leave only the specified
4608 number of files in place.
4610 * udev will now create /dev/disk/by-path links for ATA devices
4611 on kernels where that is supported.
4613 * Galician, Serbian, Turkish and Korean translations were added.
4615 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
4616 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
4617 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
4618 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
4619 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
4620 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
4621 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
4622 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
4623 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
4624 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
4625 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
4626 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
4627 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
4628 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
4629 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
4630 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
4631 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
4632 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
4634 — Berlin, 2015-10-07
4638 * The DHCP implementation of systemd-networkd gained a set of
4641 - The DHCP server now supports emitting DNS and NTP
4642 information. It may be enabled and configured via
4643 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
4644 and NTP information is enabled, but no servers are
4645 configured, the corresponding uplink information (if there
4646 is any) is propagated.
4648 - Server and client now support transmission and reception
4649 of timezone information. It can be configured via the
4650 newly introduced network options UseTimezone=,
4651 EmitTimezone=, and Timezone=. Transmission of timezone
4652 information is enabled between host and containers by
4653 default now: the container will change its local timezone
4654 to what the host has set.
4656 - Lease timeouts can now be configured via
4657 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
4659 - The DHCP server improved on the stability of
4660 leases. Clients are more likely to get the same lease
4661 information back, even if the server loses state.
4663 - The DHCP server supports two new configuration options to
4664 control the lease address pool metrics, PoolOffset= and
4667 * The encapsulation limit of tunnels in systemd-networkd may
4668 now be configured via 'EncapsulationLimit='. It allows
4669 modifying the maximum additional levels of encapsulation
4670 that are permitted to be prepended to a packet.
4672 * systemd now supports the concept of user buses replacing
4673 session buses, if used with dbus-1.10 (and enabled via dbus
4674 --enable-user-session). It previously only supported this on
4675 kdbus-enabled systems, and this release expands this to
4676 'dbus-daemon' systems.
4678 * systemd-networkd now supports predictable interface names
4681 * systemd now optionally supports the new Linux kernel
4682 "unified" control group hierarchy. If enabled via the kernel
4683 command-line option 'systemd.unified_cgroup_hierarchy=1',
4684 systemd will try to mount the unified cgroup hierarchy
4685 directly on /sys/fs/cgroup. If not enabled, or not
4686 available, systemd will fall back to the legacy cgroup
4687 hierarchy setup, as before. Host system and containers can
4688 mix and match legacy and unified hierarchies as they
4689 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
4690 environment variable to individually select the hierarchy to
4691 use for executed containers. By default, nspawn will use the
4692 unified hierarchy for the containers if the host uses the
4693 unified hierarchy, and the legacy hierarchy otherwise.
4694 Please note that at this point the unified hierarchy is an
4695 experimental kernel feature and is likely to change in one
4696 of the next kernel releases. Therefore, it should not be
4697 enabled by default in downstream distributions yet. The
4698 minimum required kernel version for the unified hierarchy to
4699 work is 4.2. Note that when the unified hierarchy is used
4700 for the first time delegated access to controllers is
4701 safe. Because of this systemd-nspawn containers will get
4702 access to controllers now, as will systemd user
4703 sessions. This means containers and user sessions may now
4704 manage their own resources, partitioning up what the system
4707 * A new special scope unit "init.scope" has been introduced
4708 that encapsulates PID 1 of the system. It may be used to
4709 determine resource usage and enforce resource limits on PID
4710 1 itself. PID 1 hence moved out of the root of the control
4713 * The cgtop tool gained support for filtering out kernel
4714 threads when counting tasks in a control group. Also, the
4715 count of processes is now recursively summed up by
4716 default. Two options -k and --recursive= have been added to
4717 revert to old behaviour. The tool has also been updated to
4718 work correctly in containers now.
4720 * systemd-nspawn's --bind= and --bind-ro= options have been
4721 extended to allow creation of non-recursive bind mounts.
4723 * libsystemd gained two new calls sd_pid_get_cgroup() and
4724 sd_peer_get_cgroup() which return the control group path of
4725 a process or peer of a connected AF_UNIX socket. This
4726 function call is particularly useful when implementing
4727 delegated subtrees support in the control group hierarchy.
4729 * The "sd-event" event loop API of libsystemd now supports
4730 correct dequeuing of real-time signals, without losing
4733 * When systemd requests a polkit decision when managing units it
4734 will now add additional fields to the request, including unit
4735 name and desired operation. This enables more powerful polkit
4736 policies, that make decisions depending on these parameters.
4738 * nspawn learnt support for .nspawn settings files, that may
4739 accompany the image files or directories of containers, and
4740 may contain additional settings for the container. This is
4741 an alternative to configuring container parameters via the
4742 nspawn command line.
4744 Contributions from: Cristian Rodríguez, Daniel Mack, David
4745 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
4746 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
4747 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
4748 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
4749 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
4750 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
4751 Andersen, Tom Gundersen, Torstein Husebø
4753 — Berlin, 2015-09-08
4757 * machinectl gained a new verb 'shell' which opens a fresh
4758 shell on the target container or the host. It is similar to
4759 the existing 'login' command of machinectl, but spawns the
4760 shell directly without prompting for username or
4761 password. The pseudo machine '.host' now refers to the local
4762 host and is used by default. Hence, 'machinectl shell' can
4763 be used as replacement for 'su -' which spawns a session as
4764 a fresh systemd unit in a way that is fully isolated from
4765 the originating session.
4767 * systemd-networkd learned to cope with private-zone DHCP
4768 options and allows other programs to query the values.
4770 * SELinux access control when enabling/disabling units is no
4771 longer enforced with this release. The previous implementation
4772 was incorrect, and a new corrected implementation is not yet
4773 available. As unit file operations are still protected via
4774 polkit and D-Bus policy this is not a security problem. Yet,
4775 distributions which care about optimal SELinux support should
4776 probably not stabilize on this release.
4778 * sd-bus gained support for matches of type "arg0has=", that
4779 test for membership of strings in string arrays sent in bus
4782 * systemd-resolved now dumps the contents of its DNS and LLMNR
4783 caches to the logs on reception of the SIGUSR1 signal. This
4784 is useful to debug DNS behaviour.
4786 * The coredumpctl tool gained a new --directory= option to
4787 operate on journal files in a specific directory.
4789 * "systemctl reboot" and related commands gained a new
4790 "--message=" option which may be used to set a free-text
4791 wall message when shutting down or rebooting the
4792 system. This message is also logged, which is useful for
4793 figuring out the reason for a reboot or shutdown a
4796 * The "systemd-resolve-host" tool's -i switch now takes
4797 network interface numbers as alternative to interface names.
4799 * A new unit file setting for services has been introduced:
4800 UtmpMode= allows configuration of how precisely systemd
4801 handles utmp and wtmp entries for the service if this is
4802 enabled. This allows writing services that appear similar to
4803 user sessions in the output of the "w", "who", "last" and
4806 * systemd-resolved will now locally synthesize DNS resource
4807 records for the "localhost" and "gateway" domains as well as
4808 the local hostname. This should ensure that clients querying
4809 RRs via resolved will get similar results as those going via
4810 NSS, if nss-myhostname is enabled.
4812 Contributions from: Alastair Hughes, Alex Crawford, Daniel
4813 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
4814 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
4815 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
4816 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
4817 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
4818 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
4819 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
4820 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
4821 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
4822 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
4823 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4825 — Berlin, 2015-08-27
4829 * The systemd-efi-boot-generator functionality was merged into
4830 systemd-gpt-auto-generator.
4832 * systemd-networkd now supports Group Policy for vxlan
4833 devices. It can be enabled via the new boolean configuration
4834 option called 'GroupPolicyExtension='.
4836 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
4837 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
4838 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
4840 — Berlin, 2015-07-31
4844 * The python-systemd code has been removed from the systemd repository.
4845 A new repository has been created which accommodates the code from
4846 now on, and we kindly ask distributions to create a separate package
4847 for this: https://github.com/systemd/python-systemd
4849 * The systemd daemon will now reload its main configuration
4850 (/etc/systemd/system.conf) on daemon-reload.
4852 * sd-dhcp now exposes vendor specific extensions via
4853 sd_dhcp_lease_get_vendor_specific().
4855 * systemd-networkd gained a number of new configuration options.
4857 - A new boolean configuration option for TAP devices called
4858 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
4859 device, thus allowing to send and receive GSO packets.
4861 - A new tunnel configuration option called 'CopyDSCP='.
4862 If enabled, the DSCP field of ip6 tunnels is copied into the
4863 decapsulated packet.
4865 - A set of boolean bridge configuration options were added.
4866 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
4867 and 'UnicastFlood=' are now parsed by networkd and applied to the
4868 respective bridge link device via the respective IFLA_BRPORT_*
4871 - A new string configuration option to override the hostname sent
4872 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
4873 is true, networkd will use the configured hostname instead of the
4874 system hostname when sending DHCP requests.
4876 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
4877 networkd will configure the IPv6 flow-label of the tunnel device
4878 according to RFC2460.
4880 - The 'macvtap' virtual network devices are now supported, similar to
4881 the already supported 'macvlan' devices.
4883 * systemd-resolved now implements RFC5452 to improve resilience against
4884 cache poisoning. Additionally, source port randomization is enabled
4885 by default to further protect against DNS spoofing attacks.
4887 * nss-mymachines now supports translating UIDs and GIDs of running
4888 containers with user-namespaces enabled. If a container 'foo'
4889 translates a host uid 'UID' to the container uid 'TUID', then
4890 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
4891 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
4892 mapped as 'vg-foo-TGID'.
4894 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
4895 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
4896 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
4897 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
4898 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
4899 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
4900 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
4901 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
4902 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
4903 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
4905 — Berlin, 2015-07-29
4909 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
4910 There are no known issues with current sysfs, and udev does not need
4911 or should be used to work around such bugs.
4913 * udev does no longer enable USB HID power management. Several reports
4914 indicate, that some devices cannot handle that setting.
4916 * The udev accelerometer helper was removed. The functionality
4917 is now fully included in iio-sensor-proxy. But this means,
4918 older iio-sensor-proxy versions will no longer provide
4919 accelerometer/orientation data with this systemd version.
4920 Please upgrade iio-sensor-proxy to version 1.0.
4922 * networkd gained a new configuration option IPv6PrivacyExtensions=
4923 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
4924 for Stateless Address") on selected networks.
4926 * For the sake of fewer build-time dependencies and less code in the
4927 main repository, the python bindings are about to be removed in the
4928 next release. A new repository has been created which accommodates
4929 the code from now on, and we kindly ask distributions to create a
4930 separate package for this. The removal will take place in v223.
4932 https://github.com/systemd/python-systemd
4934 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
4935 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
4936 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
4937 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
4938 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
4939 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
4940 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
4941 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
4942 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4943 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4945 — Berlin, 2015-07-07
4949 * The sd-bus.h and sd-event.h APIs have now been declared
4950 stable and have been added to the official interface of
4951 libsystemd.so. sd-bus implements an alternative D-Bus client
4952 library, that is relatively easy to use, very efficient and
4953 supports both classic D-Bus as well as kdbus as transport
4954 backend. sd-event is a generic event loop abstraction that
4955 is built around Linux epoll, but adds features such as event
4956 prioritization or efficient timer handling. Both APIs are good
4957 choices for C programs looking for a bus and/or event loop
4958 implementation that is minimal and does not have to be
4959 portable to other kernels.
4961 * kdbus support is no longer compile-time optional. It is now
4962 always built-in. However, it can still be disabled at
4963 runtime using the kdbus=0 kernel command line setting, and
4964 that setting may be changed to default to off, by specifying
4965 --disable-kdbus at build-time. Note though that the kernel
4966 command line setting has no effect if the kdbus.ko kernel
4967 module is not installed, in which case kdbus is (obviously)
4968 also disabled. We encourage all downstream distributions to
4969 begin testing kdbus by adding it to the kernel images in the
4970 development distributions, and leaving kdbus support in
4973 * The minimal required util-linux version has been bumped to
4976 * Support for chkconfig (--enable-chkconfig) was removed in
4977 favor of calling an abstraction tool
4978 /lib/systemd/systemd-sysv-install. This needs to be
4979 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
4980 in README for details.
4982 * If there's a systemd unit and a SysV init script for the
4983 same service name, and the user executes "systemctl enable"
4984 for it (or a related call), then this will now enable both
4985 (or execute the related operation on both), not just the
4988 * The libudev API documentation has been converted from gtkdoc
4991 * gudev has been removed from the systemd tree, it is now an
4994 * The systemd-cgtop tool learnt a new --raw switch to generate
4995 "raw" (machine parsable) output.
4997 * networkd's IPForwarding= .network file setting learnt the
4998 new setting "kernel", which ensures that networkd does not
4999 change the IP forwarding sysctl from the default kernel
5002 * The systemd-logind bus API now exposes a new boolean
5003 property "Docked" that reports whether logind considers the
5004 system "docked", i.e. connected to a docking station or not.
5006 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
5007 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
5008 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
5009 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
5010 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
5011 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
5012 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
5013 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
5014 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
5015 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
5016 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
5017 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
5018 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
5019 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
5020 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
5021 Fink, Zbigniew Jędrzejewski-Szmek
5023 — Berlin, 2015-06-19
5027 * The gudev library has been extracted into a separate repository
5028 available at: https://git.gnome.org/browse/libgudev/
5029 It is now managed as part of the Gnome project. Distributions
5030 are recommended to pass --disable-gudev to systemd and use
5031 gudev from the Gnome project instead. gudev is still included
5032 in systemd, for now. It will be removed soon, though. Please
5033 also see the announcement-thread on systemd-devel:
5034 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
5036 * systemd now exposes a CPUUsageNSec= property for each
5037 service unit on the bus, that contains the overall consumed
5038 CPU time of a service (the sum of what each process of the
5039 service consumed). This value is only available if
5040 CPUAccounting= is turned on for a service, and is then shown
5041 in the "systemctl status" output.
5043 * Support for configuring alternative mappings of the old SysV
5044 runlevels to systemd targets has been removed. They are now
5045 hardcoded in a way that runlevels 2, 3, 4 all map to
5046 multi-user.target and 5 to graphical.target (which
5047 previously was already the default behaviour).
5049 * The auto-mounter logic gained support for mount point
5050 expiry, using a new TimeoutIdleSec= setting in .automount
5051 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
5053 * The EFI System Partition (ESP) as mounted to /boot by
5054 systemd-efi-boot-generator will now be unmounted
5055 automatically after 2 minutes of not being used. This should
5056 minimize the risk of ESP corruptions.
5058 * New /etc/fstab options x-systemd.requires= and
5059 x-systemd.requires-mounts-for= are now supported to express
5060 additional dependencies for mounts. This is useful for
5061 journalling file systems that support external journal
5062 devices or overlay file systems that require underlying file
5063 systems to be mounted.
5065 * systemd does not support direct live-upgrades (via systemctl
5066 daemon-reexec) from versions older than v44 anymore. As no
5067 distribution we are aware of shipped such old versions in a
5068 stable release this should not be problematic.
5070 * When systemd forks off a new per-connection service instance
5071 it will now set the $REMOTE_ADDR environment variable to the
5072 remote IP address, and $REMOTE_PORT environment variable to
5073 the remote IP port. This behaviour is similar to the
5074 corresponding environment variables defined by CGI.
5076 * systemd-networkd gained support for uplink failure
5077 detection. The BindCarrier= option allows binding interface
5078 configuration dynamically to the link sense of other
5079 interfaces. This is useful to achieve behaviour like in
5082 * systemd-networkd gained support for configuring the DHCP
5083 client identifier to use when requesting leases.
5085 * systemd-networkd now has a per-network UseNTP= option to
5086 configure whether NTP server information acquired via DHCP
5087 is passed on to services like systemd-timesyncd.
5089 * systemd-networkd gained support for vti6 tunnels.
5091 * Note that systemd-networkd manages the sysctl variable
5092 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
5093 it is configured for since v219. The variable controls IP
5094 forwarding, and is a per-interface alternative to the global
5095 /proc/sys/net/ipv[46]/ip_forward. This setting is
5096 configurable in the IPForward= option, which defaults to
5097 "no". This means if networkd is used for an interface it is
5098 no longer sufficient to set the global sysctl option to turn
5099 on IP forwarding! Instead, the .network file option
5100 IPForward= needs to be turned on! Note that the
5101 implementation of this behaviour was broken in v219 and has
5104 * Many bonding and vxlan options are now configurable in
5107 * systemd-nspawn gained a new --property= setting to set unit
5108 properties for the container scope. This is useful for
5109 setting resource parameters (e.g. "CPUShares=500") on
5110 containers started from the command line.
5112 * systemd-nspawn gained a new --private-users= switch to make
5113 use of user namespacing available on recent Linux kernels.
5115 * systemd-nspawn may now be called as part of a shell pipeline
5116 in which case the pipes used for stdin and stdout are passed
5117 directly to the process invoked in the container, without
5118 indirection via a pseudo tty.
5120 * systemd-nspawn gained a new switch to control the UNIX
5121 signal to use when killing the init process of the container
5124 * systemd-nspawn gained a new --overlay= switch for mounting
5125 overlay file systems into the container using the new kernel
5128 * When a container image is imported via systemd-importd and
5129 the host file system is not btrfs, a loopback block device
5130 file is created in /var/lib/machines.raw with a btrfs file
5131 system inside. It is then mounted to /var/lib/machines to
5132 enable btrfs features for container management. The loopback
5133 file and btrfs file system is grown as needed when container
5134 images are imported via systemd-importd.
5136 * systemd-machined/systemd-importd gained support for btrfs
5137 quota, to enforce container disk space limits on disk. This
5138 is exposed in "machinectl set-limit".
5140 * systemd-importd now can import containers from local .tar,
5141 .raw and .qcow2 images, and export them to .tar and .raw. It
5142 can also import dkr v2 images now from the network (on top
5145 * systemd-importd gained support for verifying downloaded
5146 images with gpg2 (previously only gpg1 was supported).
5148 * systemd-machined, systemd-logind, systemd: most bus calls are
5149 now accessible to unprivileged processes via polkit. Also,
5150 systemd-logind will now allow users to kill their own sessions
5151 without further privileges or authorization.
5153 * systemd-shutdownd has been removed. This service was
5154 previously responsible for implementing scheduled shutdowns
5155 as exposed in /usr/bin/shutdown's time parameter. This
5156 functionality has now been moved into systemd-logind and is
5157 accessible via a bus interface.
5159 * "systemctl reboot" gained a new switch --firmware-setup that
5160 can be used to reboot into the EFI firmware setup, if that
5161 is available. systemd-logind now exposes an API on the bus
5162 to trigger such reboots, in case graphical desktop UIs want
5163 to cover this functionality.
5165 * "systemctl enable", "systemctl disable" and "systemctl mask"
5166 now support a new "--now" switch. If specified the units
5167 that are enabled will also be started, and the ones
5168 disabled/masked also stopped.
5170 * The Gummiboot EFI boot loader tool has been merged into
5171 systemd, and renamed to "systemd-boot". The bootctl tool has been
5172 updated to support systemd-boot.
5174 * An EFI kernel stub has been added that may be used to create
5175 kernel EFI binaries that contain not only the actual kernel,
5176 but also an initrd, boot splash, command line and OS release
5177 information. This combined binary can then be signed as a
5178 single image, so that the firmware can verify it all in one
5179 step. systemd-boot has special support for EFI binaries created
5180 like this and can extract OS release information from them
5181 and show them in the boot menu. This functionality is useful
5182 to implement cryptographically verified boot schemes.
5184 * Optional support has been added to systemd-fsck to pass
5185 fsck's progress report to an AF_UNIX socket in the file
5188 * udev will no longer create device symlinks for all block
5189 devices by default. A blacklist for excluding special block
5190 devices from this logic has been turned into a whitelist
5191 that requires picking block devices explicitly that require
5194 * A new (currently still internal) API sd-device.h has been
5195 added to libsystemd. This modernized API is supposed to
5196 replace libudev eventually. In fact, already much of libudev
5197 is now just a wrapper around sd-device.h.
5199 * A new hwdb database for storing metadata about pointing
5200 stick devices has been added.
5202 * systemd-tmpfiles gained support for setting file attributes
5203 similar to the "chattr" tool with new 'h' and 'H' lines.
5205 * systemd-journald will no longer unconditionally set the
5206 btrfs NOCOW flag on new journal files. This is instead done
5207 with tmpfiles snippet using the new 'h' line type. This
5208 allows easy disabling of this logic, by masking the
5209 journal-nocow.conf tmpfiles file.
5211 * systemd-journald will now translate audit message types to
5212 human readable identifiers when writing them to the
5213 journal. This should improve readability of audit messages.
5215 * The LUKS logic gained support for the offset= and skip=
5216 options in /etc/crypttab, as previously implemented by
5219 * /usr/lib/os-release gained a new optional field VARIANT= for
5220 distributions that support multiple variants (such as a
5221 desktop edition, a server edition, ...)
5223 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
5224 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
5225 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
5226 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
5227 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
5228 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
5229 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
5230 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
5231 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
5232 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
5233 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
5234 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
5235 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
5236 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
5237 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
5238 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
5239 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
5240 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
5241 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
5242 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
5243 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
5244 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
5245 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
5246 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
5247 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
5248 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
5249 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
5251 — Berlin, 2015-05-22
5255 * Introduce a new API "sd-hwdb.h" for querying the hardware
5256 metadata database. With this minimal interface one can query
5257 and enumerate the udev hwdb, decoupled from the old libudev
5258 library. libudev's interface for this is now only a wrapper
5259 around sd-hwdb. A new tool systemd-hwdb has been added to
5260 interface with and update the database.
5262 * When any of systemd's tools copies files (for example due to
5263 tmpfiles' C lines) a btrfs reflink will attempted first,
5264 before bytewise copying is done.
5266 * systemd-nspawn gained a new --ephemeral switch. When
5267 specified a btrfs snapshot is taken of the container's root
5268 directory, and immediately removed when the container
5269 terminates again. Thus, a container can be started whose
5270 changes never alter the container's root directory, and are
5271 lost on container termination. This switch can also be used
5272 for starting a container off the root file system of the
5273 host without affecting the host OS. This switch is only
5274 available on btrfs file systems.
5276 * systemd-nspawn gained a new --template= switch. It takes the
5277 path to a container tree to use as template for the tree
5278 specified via --directory=, should that directory be
5279 missing. This allows instantiating containers dynamically,
5280 on first run. This switch is only available on btrfs file
5283 * When a .mount unit refers to a mount point on which multiple
5284 mounts are stacked, and the .mount unit is stopped all of
5285 the stacked mount points will now be unmounted until no
5286 mount point remains.
5288 * systemd now has an explicit notion of supported and
5289 unsupported unit types. Jobs enqueued for unsupported unit
5290 types will now fail with an "unsupported" error code. More
5291 specifically .swap, .automount and .device units are not
5292 supported in containers, .busname units are not supported on
5293 non-kdbus systems. .swap and .automount are also not
5294 supported if their respective kernel compile time options
5297 * machinectl gained support for two new "copy-from" and
5298 "copy-to" commands for copying files from a running
5299 container to the host or vice versa.
5301 * machinectl gained support for a new "bind" command to bind
5302 mount host directories into local containers. This is
5303 currently only supported for nspawn containers.
5305 * networkd gained support for configuring bridge forwarding
5306 database entries (fdb) from .network files.
5308 * A new tiny daemon "systemd-importd" has been added that can
5309 download container images in tar, raw, qcow2 or dkr formats,
5310 and make them available locally in /var/lib/machines, so
5311 that they can run as nspawn containers. The daemon can GPG
5312 verify the downloads (not supported for dkr, since it has no
5313 provisions for verifying downloads). It will transparently
5314 decompress bz2, xz, gzip compressed downloads if necessary,
5315 and restore sparse files on disk. The daemon uses privilege
5316 separation to ensure the actual download logic runs with
5317 fewer privileges than the daemon itself. machinectl has
5318 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
5319 make the functionality of importd available to the
5320 user. With this in place the Fedora and Ubuntu "Cloud"
5321 images can be downloaded and booted as containers unmodified
5322 (the Fedora images lack the appropriate GPG signature files
5323 currently, so they cannot be verified, but this will change
5324 soon, hopefully). Note that downloading images is currently
5325 only fully supported on btrfs.
5327 * machinectl is now able to list container images found in
5328 /var/lib/machines, along with some metadata about sizes of
5329 disk and similar. If the directory is located on btrfs and
5330 quota is enabled, this includes quota display. A new command
5331 "image-status" has been added that shows additional
5332 information about images.
5334 * machinectl is now able to clone container images
5335 efficiently, if the underlying file system (btrfs) supports
5336 it, with the new "machinectl clone" command. It also
5337 gained commands for renaming and removing images, as well as
5338 marking them read-only or read-write (supported also on
5339 legacy file systems).
5341 * networkd gained support for collecting LLDP network
5342 announcements, from hardware that supports this. This is
5343 shown in networkctl output.
5345 * systemd-run gained support for a new -t (--pty) switch for
5346 invoking a binary on a pty whose input and output is
5347 connected to the invoking terminal. This allows executing
5348 processes as system services while interactively
5349 communicating with them via the terminal. Most interestingly
5350 this is supported across container boundaries. Invoking
5351 "systemd-run -t /bin/bash" is an alternative to running a
5352 full login session, the difference being that the former
5353 will not register a session, nor go through the PAM session
5356 * tmpfiles gained support for a new "v" line type for creating
5357 btrfs subvolumes. If the underlying file system is a legacy
5358 file system, this automatically degrades to creating a
5359 normal directory. Among others /var/lib/machines is now
5360 created like this at boot, should it be missing.
5362 * The directory /var/lib/containers/ has been deprecated and
5363 been replaced by /var/lib/machines. The term "machines" has
5364 been used in the systemd context as generic term for both
5365 VMs and containers, and hence appears more appropriate for
5366 this, as the directory can also contain raw images bootable
5369 * systemd-nspawn when invoked with -M but without --directory=
5370 or --image= is now capable of searching for the container
5371 root directory, subvolume or disk image automatically, in
5372 /var/lib/machines. systemd-nspawn@.service has been updated
5373 to make use of this, thus allowing it to be used for raw
5376 * A new machines.target unit has been introduced that is
5377 supposed to group all containers/VMs invoked as services on
5378 the system. systemd-nspawn@.service has been updated to
5379 integrate with that.
5381 * machinectl gained a new "start" command, for invoking a
5382 container as a service. "machinectl start foo" is mostly
5383 equivalent to "systemctl start systemd-nspawn@foo.service",
5384 but handles escaping in a nicer way.
5386 * systemd-nspawn will now mount most of the cgroupfs tree
5387 read-only into each container, with the exception of the
5388 container's own subtree in the name=systemd hierarchy.
5390 * journald now sets the special FS_NOCOW file flag for its
5391 journal files. This should improve performance on btrfs, by
5392 avoiding heavy fragmentation when journald's write-pattern
5393 is used on COW file systems. It degrades btrfs' data
5394 integrity guarantees for the files to the same levels as for
5395 ext3/ext4 however. This should be OK though as journald does
5396 its own data integrity checks and all its objects are
5397 checksummed on disk. Also, journald should handle btrfs disk
5398 full events a lot more gracefully now, by processing SIGBUS
5399 errors, and not relying on fallocate() anymore.
5401 * When journald detects that journal files it is writing to
5402 have been deleted it will immediately start new journal
5405 * systemd now provides a way to store file descriptors
5406 per-service in PID 1. This is useful for daemons to ensure
5407 that fds they require are not lost during a daemon
5408 restart. The fds are passed to the daemon on the next
5409 invocation in the same way socket activation fds are
5410 passed. This is now used by journald to ensure that the
5411 various sockets connected to all the system's stdout/stderr
5412 are not lost when journald is restarted. File descriptors
5413 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
5414 an extension to sd_notify(). Note that a limit is enforced
5415 on the number of fds a service can store in PID 1, and it
5416 defaults to 0, so that no fds may be stored, unless this is
5417 explicitly turned on.
5419 * The default TERM variable to use for units connected to a
5420 terminal, when no other value is explicitly is set is now
5421 vt220 rather than vt102. This should be fairly safe still,
5422 but allows PgUp/PgDn work.
5424 * The /etc/crypttab option header= as known from Debian is now
5427 * "loginctl user-status" and "loginctl session-status" will
5428 now show the last 10 lines of log messages of the
5429 user/session following the status output. Similar,
5430 "machinectl status" will show the last 10 log lines
5431 associated with a virtual machine or container
5432 service. (Note that this is usually not the log messages
5433 done in the VM/container itself, but simply what the
5434 container manager logs. For nspawn this includes all console
5437 * "loginctl session-status" without further argument will now
5438 show the status of the session of the caller. Similar,
5439 "lock-session", "unlock-session", "activate",
5440 "enable-linger", "disable-linger" may now be called without
5441 session/user parameter in which case they apply to the
5442 caller's session/user.
5444 * An X11 session scriptlet is now shipped that uploads
5445 $DISPLAY and $XAUTHORITY into the environment of the systemd
5446 --user daemon if a session begins. This should improve
5447 compatibility with X11 enabled applications run as systemd
5450 * Generators are now subject to masking via /etc and /run, the
5451 same way as unit files.
5453 * networkd .network files gained support for configuring
5454 per-link IPv4/IPv6 packet forwarding as well as IPv4
5455 masquerading. This is by default turned on for veth links to
5456 containers, as registered by systemd-nspawn. This means that
5457 nspawn containers run with --network-veth will now get
5458 automatic routed access to the host's networks without any
5459 further configuration or setup, as long as networkd runs on
5462 * systemd-nspawn gained the --port= (-p) switch to expose TCP
5463 or UDP posts of a container on the host. With this in place
5464 it is possible to run containers with private veth links
5465 (--network-veth), and have their functionality exposed on
5466 the host as if their services were running directly on the
5469 * systemd-nspawn's --network-veth switch now gained a short
5470 version "-n", since with the changes above it is now truly
5471 useful out-of-the-box. The systemd-nspawn@.service has been
5472 updated to make use of it too by default.
5474 * systemd-nspawn will now maintain a per-image R/W lock, to
5475 ensure that the same image is not started more than once
5476 writable. (It's OK to run an image multiple times
5477 simultaneously in read-only mode.)
5479 * systemd-nspawn's --image= option is now capable of
5480 dissecting and booting MBR and GPT disk images that contain
5481 only a single active Linux partition. Previously it
5482 supported only GPT disk images with proper GPT type
5483 IDs. This allows running cloud images from major
5484 distributions directly with systemd-nspawn, without
5487 * In addition to collecting mouse dpi data in the udev
5488 hardware database, there's now support for collecting angle
5489 information for mouse scroll wheels. The database is
5490 supposed to guarantee similar scrolling behavior on mice
5491 that it knows about. There's also support for collecting
5492 information about Touchpad types.
5494 * udev's input_id built-in will now also collect touch screen
5495 dimension data and attach it to probed devices.
5497 * /etc/os-release gained support for a Distribution Privacy
5500 * networkd gained support for creating "ipvlan", "gretap",
5501 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
5503 * systemd-tmpfiles gained support for "a" lines for setting
5506 * systemd-nspawn will now mount /tmp in the container to
5507 tmpfs, automatically.
5509 * systemd now exposes the memory.usage_in_bytes cgroup
5510 attribute and shows it for each service in the "systemctl
5511 status" output, if available.
5513 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
5514 immediate reboot is triggered. This useful if shutdown is
5515 hung and is unable to complete, to expedite the
5516 operation. Note that this kind of reboot will still unmount
5517 all file systems, and hence should not result in fsck being
5520 * A .device unit for an optical block device will now be
5521 considered active only when a medium is in the drive. Also,
5522 mount units are now bound to their backing devices thus
5523 triggering automatic unmounting when devices become
5524 unavailable. With this in place systemd will now
5525 automatically unmount left-over mounts when a CD-ROM is
5526 ejected or an USB stick is yanked from the system.
5528 * networkd-wait-online now has support for waiting for
5529 specific interfaces only (with globbing), and for giving up
5530 after a configurable timeout.
5532 * networkd now exits when idle. It will be automatically
5533 restarted as soon as interfaces show up, are removed or
5534 change state. networkd will stay around as long as there is
5535 at least one DHCP state machine or similar around, that keep
5538 * networkd may now configure IPv6 link-local addressing in
5539 addition to IPv4 link-local addressing.
5541 * The IPv6 "token" for use in SLAAC may now be configured for
5542 each .network interface in networkd.
5544 * Routes configured with networkd may now be assigned a scope
5547 * networkd's [Match] sections now support globbing and lists
5548 of multiple space-separated matches per item.
5550 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
5551 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
5552 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
5553 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
5554 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
5555 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
5556 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
5557 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
5558 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
5559 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
5560 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
5561 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
5562 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
5563 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
5564 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
5565 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
5566 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
5567 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
5568 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
5569 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
5570 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
5571 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
5572 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
5573 Hoffmann, Zbigniew Jędrzejewski-Szmek
5575 — Berlin, 2015-02-16
5579 * When querying unit file enablement status (for example via
5580 "systemctl is-enabled"), a new state "indirect" is now known
5581 which indicates that a unit might not be enabled itself, but
5582 another unit listed in its Also= setting might be.
5584 * Similar to the various existing ConditionXYZ= settings for
5585 units, there are now matching AssertXYZ= settings. While
5586 failing conditions cause a unit to be skipped, but its job
5587 to succeed, failing assertions declared like this will cause
5588 a unit start operation and its job to fail.
5590 * hostnamed now knows a new chassis type "embedded".
5592 * systemctl gained a new "edit" command. When used on a unit
5593 file, this allows extending unit files with .d/ drop-in
5594 configuration snippets or editing the full file (after
5595 copying it from /usr/lib to /etc). This will invoke the
5596 user's editor (as configured with $EDITOR), and reload the
5597 modified configuration after editing.
5599 * "systemctl status" now shows the suggested enablement state
5600 for a unit, as declared in the (usually vendor-supplied)
5601 system preset files.
5603 * nss-myhostname will now resolve the single-label host name
5604 "gateway" to the locally configured default IP routing
5605 gateways, ordered by their metrics. This assigns a stable
5606 name to the used gateways, regardless which ones are
5607 currently configured. Note that the name will only be
5608 resolved after all other name sources (if nss-myhostname is
5609 configured properly) and should hence not negatively impact
5610 systems that use the single-label host name "gateway" in
5613 * systemd-inhibit now allows filtering by mode when listing
5616 * Scope and service units gained a new "Delegate" boolean
5617 property, which, when set, allows processes running inside the
5618 unit to further partition resources. This is primarily
5619 useful for systemd user instances as well as container
5622 * journald will now pick up audit messages directly from
5623 the kernel, and log them like any other log message. The
5624 audit fields are split up and fully indexed. This means that
5625 journalctl in many ways is now a (nicer!) alternative to
5626 ausearch, the traditional audit client. Note that this
5627 implements only a minimal audit client. If you want the
5628 special audit modes like reboot-on-log-overflow, please use
5629 the traditional auditd instead, which can be used in
5630 parallel to journald.
5632 * The ConditionSecurity= unit file option now understands the
5633 special string "audit" to check whether auditing is
5636 * journalctl gained two new commands --vacuum-size= and
5637 --vacuum-time= to delete old journal files until the
5638 remaining ones take up no more than the specified size on disk,
5639 or are not older than the specified time.
5641 * A new, native PPPoE library has been added to sd-network,
5642 systemd's library of light-weight networking protocols. This
5643 library will be used in a future version of networkd to
5644 enable PPPoE communication without an external pppd daemon.
5646 * The busctl tool now understands a new "capture" verb that
5647 works similar to "monitor", but writes a packet capture
5648 trace to STDOUT that can be redirected to a file which is
5649 compatible with libcap's capture file format. This can then
5650 be loaded in Wireshark and similar tools to inspect bus
5653 * The busctl tool now understands a new "tree" verb that shows
5654 the object trees of a specific service on the bus, or of all
5657 * The busctl tool now understands a new "introspect" verb that
5658 shows all interfaces and members of objects on the bus,
5659 including their signature and values. This is particularly
5660 useful to get more information about bus objects shown by
5661 the new "busctl tree" command.
5663 * The busctl tool now understands new verbs "call",
5664 "set-property" and "get-property" for invoking bus method
5665 calls, setting and getting bus object properties in a
5668 * busctl gained a new --augment-creds= argument that controls
5669 whether the tool shall augment credential information it
5670 gets from the bus with data from /proc, in a possibly
5673 * nspawn's --link-journal= switch gained two new values
5674 "try-guest" and "try-host" that work like "guest" and
5675 "host", but do not fail if the host has no persistent
5676 journalling enabled. -j is now equivalent to
5677 --link-journal=try-guest.
5679 * macvlan network devices created by nspawn will now have
5680 stable MAC addresses.
5682 * A new SmackProcessLabel= unit setting has been added, which
5683 controls the SMACK security label processes forked off by
5684 the respective unit shall use.
5686 * If compiled with --enable-xkbcommon, systemd-localed will
5687 verify x11 keymap settings by compiling the given keymap. It
5688 will spew out warnings if the compilation fails. This
5689 requires libxkbcommon to be installed.
5691 * When a coredump is collected, a larger number of metadata
5692 fields is now collected and included in the journal records
5693 created for it. More specifically, control group membership,
5694 environment variables, memory maps, working directory,
5695 chroot directory, /proc/$PID/status, and a list of open file
5696 descriptors is now stored in the log entry.
5698 * The udev hwdb now contains DPI information for mice. For
5701 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
5703 * All systemd programs that read standalone configuration
5704 files in /etc now also support a corresponding series of
5705 .conf.d configuration directories in /etc/, /run/,
5706 /usr/local/lib/, /usr/lib/, and (if configured with
5707 --enable-split-usr) /lib/. In particular, the following
5708 configuration files now have corresponding configuration
5709 directories: system.conf user.conf, logind.conf,
5710 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
5711 resolved.conf, timesyncd.conf, journal-remote.conf, and
5712 journal-upload.conf. Note that distributions should use the
5713 configuration directories in /usr/lib/; the directories in
5714 /etc/ are reserved for the system administrator.
5716 * systemd-rfkill will no longer take the rfkill device name
5717 into account when storing rfkill state on disk, as the name
5718 might be dynamically assigned and not stable. Instead, the
5719 ID_PATH udev variable combined with the rfkill type (wlan,
5720 bluetooth, ...) is used.
5722 * A new service systemd-machine-id-commit.service has been
5723 added. When used on systems where /etc is read-only during
5724 boot, and /etc/machine-id is not initialized (but an empty
5725 file), this service will copy the temporary machine ID
5726 created as replacement into /etc after the system is fully
5727 booted up. This is useful for systems that are freshly
5728 installed with a non-initialized machine ID, but should get
5729 a fixed machine ID for subsequent boots.
5731 * networkd's .netdev files now provide a large set of
5732 configuration parameters for VXLAN devices. Similarly, the
5733 bridge port cost parameter is now configurable in .network
5734 files. There's also new support for configuring IP source
5735 routing. networkd .link files gained support for a new
5736 OriginalName= match that is useful to match against the
5737 original interface name the kernel assigned. .network files
5738 may include MTU= and MACAddress= fields for altering the MTU
5739 and MAC address while being connected to a specific network
5742 * The LUKS logic gained supported for configuring
5743 UUID-specific key files. There's also new support for naming
5744 LUKS device from the kernel command line, using the new
5745 luks.name= argument.
5747 * Timer units may now be transiently created via the bus API
5748 (this was previously already available for scope and service
5749 units). In addition it is now possible to create multiple
5750 transient units at the same time with a single bus call. The
5751 "systemd-run" tool has been updated to make use of this for
5752 running commands on a specified time, in at(1)-style.
5754 * tmpfiles gained support for "t" lines, for assigning
5755 extended attributes to files. Among other uses this may be
5756 used to assign SMACK labels to files.
5758 Contributions from: Alin Rauta, Alison Chaiken, Andrej
5759 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
5760 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
5761 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
5762 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
5763 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
5764 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
5765 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
5766 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
5767 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
5768 Michael Biebl, Michael Chapman, Michael Marineau, Michal
5769 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
5770 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
5771 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
5772 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
5773 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5774 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
5775 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
5777 — Berlin, 2014-12-10
5781 * journalctl gained the new options -t/--identifier= to match
5782 on the syslog identifier (aka "tag"), as well as --utc to
5783 show log timestamps in the UTC timezone. journalctl now also
5784 accepts -n/--lines=all to disable line capping in a pager.
5786 * journalctl gained a new switch, --flush, that synchronously
5787 flushes logs from /run/log/journal to /var/log/journal if
5788 persistent storage is enabled. systemd-journal-flush.service
5789 now waits until the operation is complete.
5791 * Services can notify the manager before they start a reload
5792 (by sending RELOADING=1) or shutdown (by sending
5793 STOPPING=1). This allows the manager to track and show the
5794 internal state of daemons and closes a race condition when
5795 the process is still running but has closed its D-Bus
5798 * Services with Type=oneshot do not have to have any ExecStart
5801 * User units are now loaded also from
5802 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
5803 /run/systemd/user directory that was already previously
5804 supported, but is under the control of the user.
5806 * Job timeouts (i.e. timeouts on the time a job that is
5807 queued stays in the run queue) can now optionally result in
5808 immediate reboot or power-off actions (JobTimeoutAction= and
5809 JobTimeoutRebootArgument=). This is useful on ".target"
5810 units, to limit the maximum time a target remains
5811 undispatched in the run queue, and to trigger an emergency
5812 operation in such a case. This is now used by default to
5813 turn off the system if boot-up (as defined by everything in
5814 basic.target) hangs and does not complete for at least
5815 15min. Also, if power-off or reboot hang for at least 30min
5816 an immediate power-off/reboot operation is triggered. This
5817 functionality is particularly useful to increase reliability
5818 on embedded devices, but also on laptops which might
5819 accidentally get powered on when carried in a backpack and
5820 whose boot stays stuck in a hard disk encryption passphrase
5823 * systemd-logind can be configured to also handle lid switch
5824 events even when the machine is docked or multiple displays
5825 are attached (HandleLidSwitchDocked= option).
5827 * A helper binary and a service have been added which can be
5828 used to resume from hibernation in the initramfs. A
5829 generator will parse the resume= option on the kernel
5830 command line to trigger resume.
5832 * A user console daemon systemd-consoled has been
5833 added. Currently, it is a preview, and will so far open a
5834 single terminal on each session of the user marked as
5835 Desktop=systemd-console.
5837 * Route metrics can be specified for DHCP routes added by
5840 * The SELinux context of socket-activated services can be set
5841 from the information provided by the networking stack
5842 (SELinuxContextFromNet= option).
5844 * Userspace firmware loading support has been removed and
5845 the minimum supported kernel version is thus bumped to 3.7.
5847 * Timeout for udev workers has been increased from 1 to 3
5848 minutes, but a warning will be printed after 1 minute to
5849 help diagnose kernel modules that take a long time to load.
5851 * Udev rules can now remove tags on devices with TAG-="foobar".
5853 * systemd's readahead implementation has been removed. In many
5854 circumstances it didn't give expected benefits even for
5855 rotational disk drives and was becoming less relevant in the
5856 age of SSDs. As none of the developers has been using
5857 rotating media anymore, and nobody stepped up to actively
5858 maintain this component of systemd it has now been removed.
5860 * Swap units can use Options= to specify discard options.
5861 Discard options specified for swaps in /etc/fstab are now
5864 * Docker containers are now detected as a separate type of
5867 * The Password Agent protocol gained support for queries where
5868 the user input is shown, useful e.g. for user names.
5869 systemd-ask-password gained a new --echo option to turn that
5872 * The default sysctl.d/ snippets will now set:
5874 net.core.default_qdisc = fq_codel
5876 This selects Fair Queuing Controlled Delay as the default
5877 queuing discipline for network interfaces. fq_codel helps
5878 fight the network bufferbloat problem. It is believed to be
5879 a good default with no tuning required for most workloads.
5880 Downstream distributions may override this choice. On 10Gbit
5881 servers that do not do forwarding, "fq" may perform better.
5882 Systems without a good clocksource should use "pfifo_fast".
5884 * If kdbus is enabled during build a new option BusPolicy= is
5885 available for service units, that allows locking all service
5886 processes into a stricter bus policy, in order to limit
5887 access to various bus services, or even hide most of them
5888 from the service's view entirely.
5890 * networkctl will now show the .network and .link file
5891 networkd has applied to a specific interface.
5893 * sd-login gained a new API call sd_session_get_desktop() to
5894 query which desktop environment has been selected for a
5897 * UNIX utmp support is now compile-time optional to support
5898 legacy-free systems.
5900 * systemctl gained two new commands "add-wants" and
5901 "add-requires" for pulling in units from specific targets
5904 * If the word "rescue" is specified on the kernel command line
5905 the system will now boot into rescue mode (aka
5906 rescue.target), which was previously available only by
5907 specifying "1" or "systemd.unit=rescue.target" on the kernel
5908 command line. This new kernel command line option nicely
5909 mirrors the already existing "emergency" kernel command line
5912 * New kernel command line options mount.usr=, mount.usrflags=,
5913 mount.usrfstype= have been added that match root=, rootflags=,
5914 rootfstype= but allow mounting a specific file system to
5917 * The $NOTIFY_SOCKET is now also passed to control processes of
5918 services, not only the main process.
5920 * This version reenables support for fsck's -l switch. This
5921 means at least version v2.25 of util-linux is required for
5922 operation, otherwise dead-locks on device nodes may
5923 occur. Again: you need to update util-linux to at least
5924 v2.25 when updating systemd to v217.
5926 * The "multi-seat-x" tool has been removed from systemd, as
5927 its functionality has been integrated into X servers 1.16,
5928 and the tool is hence redundant. It is recommended to update
5929 display managers invoking this tool to simply invoke X
5930 directly from now on, again.
5932 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
5933 message flag has been added for all of systemd's polkit
5934 authenticated method calls has been added. In particular this
5935 now allows optional interactive authorization via polkit for
5936 many of PID1's privileged operations such as unit file
5937 enabling and disabling.
5939 * "udevadm hwdb --update" learnt a new switch "--usr" for
5940 placing the rebuilt hardware database in /usr instead of
5941 /etc. When used only hardware database entries stored in
5942 /usr will be used, and any user database entries in /etc are
5943 ignored. This functionality is useful for vendors to ship a
5944 pre-built database on systems where local configuration is
5945 unnecessary or unlikely.
5947 * Calendar time specifications in .timer units now also
5948 understand the strings "semi-annually", "quarterly" and
5949 "minutely" as shortcuts (in addition to the preexisting
5950 "anually", "hourly", ...).
5952 * systemd-tmpfiles will now correctly create files in /dev
5953 at boot which are marked for creation only at boot. It is
5954 recommended to always create static device nodes with 'c!'
5955 and 'b!', so that they are created only at boot and not
5956 overwritten at runtime.
5958 * When the watchdog logic is used for a service (WatchdogSec=)
5959 and the watchdog timeout is hit the service will now be
5960 terminated with SIGABRT (instead of just SIGTERM), in order
5961 to make sure a proper coredump and backtrace is
5962 generated. This ensures that hanging services will result in
5963 similar coredump/backtrace behaviour as services that hit a
5966 Contributions from: Andreas Henriksson, Andrei Borzenkov,
5967 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
5968 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
5969 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
5970 Herrmann, David Sommerseth, David Strauss, Emil Renner
5971 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
5972 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
5973 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
5974 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
5975 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
5976 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
5977 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
5978 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
5979 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
5980 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
5981 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
5982 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
5983 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
5984 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
5985 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
5986 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
5989 — Berlin, 2014-10-28
5993 * timedated no longer reads NTP implementation unit names from
5994 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
5995 implementations should add a
5997 Conflicts=systemd-timesyncd.service
5999 to their unit files to take over and replace systemd's NTP
6000 default functionality.
6002 * systemd-sysusers gained a new line type "r" for configuring
6003 which UID/GID ranges to allocate system users/groups
6004 from. Lines of type "u" may now add an additional column
6005 that specifies the home directory for the system user to be
6006 created. Also, systemd-sysusers may now optionally read user
6007 information from STDIN instead of a file. This is useful for
6008 invoking it from RPM preinst scriptlets that need to create
6009 users before the first RPM file is installed since these
6010 files might need to be owned by them. A new
6011 %sysusers_create_inline RPM macro has been introduced to do
6012 just that. systemd-sysusers now updates the shadow files as
6013 well as the user/group databases, which should enhance
6014 compatibility with certain tools like grpck.
6016 * A number of bus APIs of PID 1 now optionally consult polkit to
6017 permit access for otherwise unprivileged clients under certain
6018 conditions. Note that this currently doesn't support
6019 interactive authentication yet, but this is expected to be
6020 added eventually, too.
6022 * /etc/machine-info now has new fields for configuring the
6023 deployment environment of the machine, as well as the
6024 location of the machine. hostnamectl has been updated with
6025 new command to update these fields.
6027 * systemd-timesyncd has been updated to automatically acquire
6028 NTP server information from systemd-networkd, which might
6029 have been discovered via DHCP.
6031 * systemd-resolved now includes a caching DNS stub resolver
6032 and a complete LLMNR name resolution implementation. A new
6033 NSS module "nss-resolve" has been added which can be used
6034 instead of glibc's own "nss-dns" to resolve hostnames via
6035 systemd-resolved. Hostnames, addresses and arbitrary RRs may
6036 be resolved via systemd-resolved D-Bus APIs. In contrast to
6037 the glibc internal resolver systemd-resolved is aware of
6038 multi-homed system, and keeps DNS server and caches separate
6039 and per-interface. Queries are sent simultaneously on all
6040 interfaces that have DNS servers configured, in order to
6041 properly handle VPNs and local LANs which might resolve
6042 separate sets of domain names. systemd-resolved may acquire
6043 DNS server information from systemd-networkd automatically,
6044 which in turn might have discovered them via DHCP. A tool
6045 "systemd-resolve-host" has been added that may be used to
6046 query the DNS logic in resolved. systemd-resolved implements
6047 IDNA and automatically uses IDNA or UTF-8 encoding depending
6048 on whether classic DNS or LLMNR is used as transport. In the
6049 next releases we intend to add a DNSSEC and mDNS/DNS-SD
6050 implementation to systemd-resolved.
6052 * A new NSS module nss-mymachines has been added, that
6053 automatically resolves the names of all local registered
6054 containers to their respective IP addresses.
6056 * A new client tool "networkctl" for systemd-networkd has been
6057 added. It currently is entirely passive and will query
6058 networking configuration from udev, rtnetlink and networkd,
6059 and present it to the user in a very friendly
6060 way. Eventually, we hope to extend it to become a full
6061 control utility for networkd.
6063 * .socket units gained a new DeferAcceptSec= setting that
6064 controls the kernels' TCP_DEFER_ACCEPT sockopt for
6065 TCP. Similarly, support for controlling TCP keep-alive
6066 settings has been added (KeepAliveTimeSec=,
6067 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
6068 turning off Nagle's algorithm on TCP has been added
6071 * logind learned a new session type "web", for use in projects
6072 like Cockpit which register web clients as PAM sessions.
6074 * timer units with at least one OnCalendar= setting will now
6075 be started only after time-sync.target has been
6076 reached. This way they will not elapse before the system
6077 clock has been corrected by a local NTP client or
6078 similar. This is particular useful on RTC-less embedded
6079 machines, that come up with an invalid system clock.
6081 * systemd-nspawn's --network-veth= switch should now result in
6082 stable MAC addresses for both the outer and the inner side
6085 * systemd-nspawn gained a new --volatile= switch for running
6086 container instances with /etc or /var unpopulated.
6088 * The kdbus client code has been updated to use the new Linux
6089 3.17 memfd subsystem instead of the old kdbus-specific one.
6091 * systemd-networkd's DHCP client and server now support
6092 FORCERENEW. There are also new configuration options to
6093 configure the vendor client identifier and broadcast mode
6096 * systemd will no longer inform the kernel about the current
6097 timezone, as this is necessarily incorrect and racy as the
6098 kernel has no understanding of DST and similar
6099 concepts. This hence means FAT timestamps will be always
6100 considered UTC, similar to what Android is already
6101 doing. Also, when the RTC is configured to the local time
6102 (rather than UTC) systemd will never synchronize back to it,
6103 as this might confuse Windows at a later boot.
6105 * systemd-analyze gained a new command "verify" for offline
6106 validation of unit files.
6108 * systemd-networkd gained support for a couple of additional
6109 settings for bonding networking setups. Also, the metric for
6110 statically configured routes may now be configured. For
6111 network interfaces where this is appropriate the peer IP
6112 address may now be configured.
6114 * systemd-networkd's DHCP client will no longer request
6115 broadcasting by default, as this tripped up some networks.
6116 For hardware where broadcast is required the feature should
6117 be switched back on using RequestBroadcast=yes.
6119 * systemd-networkd will now set up IPv4LL addresses (when
6120 enabled) even if DHCP is configured successfully.
6122 * udev will now default to respect network device names given
6123 by the kernel when the kernel indicates that these are
6124 predictable. This behavior can be tweaked by changing
6125 NamePolicy= in the relevant .link file.
6127 * A new library systemd-terminal has been added that
6128 implements full TTY stream parsing and rendering. This
6129 library is supposed to be used later on for implementing a
6130 full userspace VT subsystem, replacing the current kernel
6133 * A new tool systemd-journal-upload has been added to push
6134 journal data to a remote system running
6135 systemd-journal-remote.
6137 * journald will no longer forward all local data to another
6138 running syslog daemon. This change has been made because
6139 rsyslog (which appears to be the most commonly used syslog
6140 implementation these days) no longer makes use of this, and
6141 instead pulls the data out of the journal on its own. Since
6142 forwarding the messages to a non-existent syslog server is
6143 more expensive than we assumed we have now turned this
6144 off. If you run a syslog server that is not a recent rsyslog
6145 version, you have to turn this option on again
6146 (ForwardToSyslog= in journald.conf).
6148 * journald now optionally supports the LZ4 compressor for
6149 larger journal fields. This compressor should perform much
6150 better than XZ which was the previous default.
6152 * machinectl now shows the IP addresses of local containers,
6153 if it knows them, plus the interface name of the container.
6155 * A new tool "systemd-escape" has been added that makes it
6156 easy to escape strings to build unit names and similar.
6158 * sd_notify() messages may now include a new ERRNO= field
6159 which is parsed and collected by systemd and shown among the
6160 "systemctl status" output for a service.
6162 * A new component "systemd-firstboot" has been added that
6163 queries the most basic systemd information (timezone,
6164 hostname, root password) interactively on first
6165 boot. Alternatively it may also be used to provision these
6166 things offline on OS images installed into directories.
6168 * The default sysctl.d/ snippets will now set
6170 net.ipv4.conf.default.promote_secondaries=1
6172 This has the benefit of no flushing secondary IP addresses
6173 when primary addresses are removed.
6175 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
6176 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
6177 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
6178 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
6179 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
6180 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
6181 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
6182 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
6183 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
6184 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
6185 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
6186 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
6187 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
6188 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
6189 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
6191 — Berlin, 2014-08-19
6195 * A new tool systemd-sysusers has been added. This tool
6196 creates system users and groups in /etc/passwd and
6197 /etc/group, based on static declarative system user/group
6198 definitions in /usr/lib/sysusers.d/. This is useful to
6199 enable factory resets and volatile systems that boot up with
6200 an empty /etc directory, and thus need system users and
6201 groups created during early boot. systemd now also ships
6202 with two default sysusers.d/ files for the most basic
6203 users and groups systemd and the core operating system
6206 * A new tmpfiles snippet has been added that rebuilds the
6207 essential files in /etc on boot, should they be missing.
6209 * A directive for ensuring automatic clean-up of
6210 /var/cache/man/ has been removed from the default
6211 configuration. This line should now be shipped by the man
6212 implementation. The necessary change has been made to the
6213 man-db implementation. Note that you need to update your man
6214 implementation to one that ships this line, otherwise no
6215 automatic clean-up of /var/cache/man will take place.
6217 * A new condition ConditionNeedsUpdate= has been added that
6218 may conditionalize services to only run when /etc or /var
6219 are "older" than the vendor operating system resources in
6220 /usr. This is useful for reconstructing or updating /etc
6221 after an offline update of /usr or a factory reset, on the
6222 next reboot. Services that want to run once after such an
6223 update or reset should use this condition and order
6224 themselves before the new systemd-update-done.service, which
6225 will mark the two directories as fully updated. A number of
6226 service files have been added making use of this, to rebuild
6227 the udev hardware database, the journald message catalog and
6228 dynamic loader cache (ldconfig). The systemd-sysusers tool
6229 described above also makes use of this now. With this in
6230 place it is now possible to start up a minimal operating
6231 system with /etc empty cleanly. For more information on the
6232 concepts involved see this recent blog story:
6234 http://0pointer.de/blog/projects/stateless.html
6236 * A new system group "input" has been introduced, and all
6237 input device nodes get this group assigned. This is useful
6238 for system-level software to get access to input devices. It
6239 complements what is already done for "audio" and "video".
6241 * systemd-networkd learnt minimal DHCPv4 server support in
6242 addition to the existing DHCPv4 client support. It also
6243 learnt DHCPv6 client and IPv6 Router Solicitation client
6244 support. The DHCPv4 client gained support for static routes
6245 passed in from the server. Note that the [DHCPv4] section
6246 known in older systemd-networkd versions has been renamed to
6247 [DHCP] and is now also used by the DHCPv6 client. Existing
6248 .network files using settings of this section should be
6249 updated, though compatibility is maintained. Optionally, the
6250 client hostname may now be sent to the DHCP server.
6252 * networkd gained support for vxlan virtual networks as well
6253 as tun/tap and dummy devices.
6255 * networkd gained support for automatic allocation of address
6256 ranges for interfaces from a system-wide pool of
6257 addresses. This is useful for dynamically managing a large
6258 number of interfaces with a single network configuration
6259 file. In particular this is useful to easily assign
6260 appropriate IP addresses to the veth links of a large number
6261 of nspawn instances.
6263 * RPM macros for processing sysusers, sysctl and binfmt
6264 drop-in snippets at package installation time have been
6267 * The /etc/os-release file should now be placed in
6268 /usr/lib/os-release. The old location is automatically
6269 created as symlink. /usr/lib is the more appropriate
6270 location of this file, since it shall actually describe the
6271 vendor operating system shipped in /usr, and not the
6272 configuration stored in /etc.
6274 * .mount units gained a new boolean SloppyOptions= setting
6275 that maps to mount(8)'s -s option which enables permissive
6276 parsing of unknown mount options.
6278 * tmpfiles learnt a new "L+" directive which creates a symlink
6279 but (unlike "L") deletes a pre-existing file first, should
6280 it already exist and not already be the correct
6281 symlink. Similarly, "b+", "c+" and "p+" directives have been
6282 added as well, which create block and character devices, as
6283 well as fifos in the filesystem, possibly removing any
6284 pre-existing files of different types.
6286 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
6287 'argument' field (which so far specified the source to
6288 symlink/copy the files from) is now optional. If omitted the
6289 same file os copied from /usr/share/factory/ suffixed by the
6290 full destination path. This is useful for populating /etc
6291 with essential files, by copying them from vendor defaults
6292 shipped in /usr/share/factory/etc.
6294 * A new command "systemctl preset-all" has been added that
6295 applies the service preset settings to all installed unit
6296 files. A new switch --preset-mode= has been added that
6297 controls whether only enable or only disable operations
6300 * A new command "systemctl is-system-running" has been added
6301 that allows checking the overall state of the system, for
6302 example whether it is fully up and running.
6304 * When the system boots up with an empty /etc, the equivalent
6305 to "systemctl preset-all" is executed during early boot, to
6306 make sure all default services are enabled after a factory
6309 * systemd now contains a minimal preset file that enables the
6310 most basic services systemd ships by default.
6312 * Unit files' [Install] section gained a new DefaultInstance=
6313 field for defining the default instance to create if a
6314 template unit is enabled with no instance specified.
6316 * A new passive target cryptsetup-pre.target has been added
6317 that may be used by services that need to make they run and
6318 finish before the first LUKS cryptographic device is set up.
6320 * The /dev/loop-control and /dev/btrfs-control device nodes
6321 are now owned by the "disk" group by default, opening up
6322 access to this group.
6324 * systemd-coredump will now automatically generate a
6325 stack trace of all core dumps taking place on the system,
6326 based on elfutils' libdw library. This stack trace is logged
6329 * systemd-coredump may now optionally store coredumps directly
6330 on disk (in /var/lib/systemd/coredump, possibly compressed),
6331 instead of storing them unconditionally in the journal. This
6332 mode is the new default. A new configuration file
6333 /etc/systemd/coredump.conf has been added to configure this
6334 and other parameters of systemd-coredump.
6336 * coredumpctl gained a new "info" verb to show details about a
6337 specific coredump. A new switch "-1" has also been added
6338 that makes sure to only show information about the most
6339 recent entry instead of all entries. Also, as the tool is
6340 generally useful now the "systemd-" prefix of the binary
6341 name has been removed. Distributions that want to maintain
6342 compatibility with the old name should add a symlink from
6343 the old name to the new name.
6345 * journald's SplitMode= now defaults to "uid". This makes sure
6346 that unprivileged users can access their own coredumps with
6347 coredumpctl without restrictions.
6349 * New kernel command line options "systemd.wants=" (for
6350 pulling an additional unit during boot), "systemd.mask="
6351 (for masking a specific unit for the boot), and
6352 "systemd.debug-shell" (for enabling the debug shell on tty9)
6353 have been added. This is implemented in the new generator
6354 "systemd-debug-generator".
6356 * systemd-nspawn will now by default filter a couple of
6357 syscalls for containers, among them those required for
6358 kernel module loading, direct x86 IO port access, swap
6359 management, and kexec. Most importantly though
6360 open_by_handle_at() is now prohibited for containers,
6361 closing a hole similar to a recently discussed vulnerability
6362 in docker regarding access to files on file hierarchies the
6363 container should normally not have access to. Note that, for
6364 nspawn, we generally make no security claims anyway (and
6365 this is explicitly documented in the man page), so this is
6366 just a fix for one of the most obvious problems.
6368 * A new man page file-hierarchy(7) has been added that
6369 contains a minimized, modernized version of the file system
6370 layout systemd expects, similar in style to the FHS
6371 specification or hier(5). A new tool systemd-path(1) has
6372 been added to query many of these paths for the local
6375 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
6376 longer done. Since the directory now has a per-user size
6377 limit, and is cleaned on logout this appears unnecessary,
6378 in particular since this now brings the lifecycle of this
6379 directory closer in line with how IPC objects are handled.
6381 * systemd.pc now exports a number of additional directories,
6382 including $libdir (which is useful to identify the library
6383 path for the primary architecture of the system), and a
6384 couple of drop-in directories.
6386 * udev's predictable network interface names now use the dev_port
6387 sysfs attribute, introduced in linux 3.15 instead of dev_id to
6388 distinguish between ports of the same PCI function. dev_id should
6389 only be used for ports using the same HW address, hence the need
6392 * machined has been updated to export the OS version of a
6393 container (read from /etc/os-release and
6394 /usr/lib/os-release) on the bus. This is now shown in
6395 "machinectl status" for a machine.
6397 * A new service setting RestartForceExitStatus= has been
6398 added. If configured to a set of exit signals or process
6399 return values, the service will be restarted when the main
6400 daemon process exits with any of them, regardless of the
6403 * systemctl's -H switch for connecting to remote systemd
6404 machines has been extended so that it may be used to
6405 directly connect to a specific container on the
6406 host. "systemctl -H root@foobar:waldi" will now connect as
6407 user "root" to host "foobar", and then proceed directly to
6408 the container named "waldi". Note that currently you have to
6409 authenticate as user "root" for this to work, as entering
6410 containers is a privileged operation.
6412 Contributions from: Andreas Henriksson, Benjamin Steinwender,
6413 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
6414 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
6415 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
6416 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
6417 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
6418 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
6419 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
6420 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
6421 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
6422 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
6423 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
6425 — Berlin, 2014-07-03
6429 * As an experimental feature, udev now tries to lock the
6430 disk device node (flock(LOCK_SH|LOCK_NB)) while it
6431 executes events for the disk or any of its partitions.
6432 Applications like partitioning programs can lock the
6433 disk device node (flock(LOCK_EX)) and claim temporary
6434 device ownership that way; udev will entirely skip all event
6435 handling for this disk and its partitions. If the disk
6436 was opened for writing, the close will trigger a partition
6437 table rescan in udev's "watch" facility, and if needed
6438 synthesize "change" events for the disk and all its partitions.
6439 This is now unconditionally enabled, and if it turns out to
6440 cause major problems, we might turn it on only for specific
6441 devices, or might need to disable it entirely. Device Mapper
6442 devices are excluded from this logic.
6444 * We temporarily dropped the "-l" switch for fsck invocations,
6445 since they collide with the flock() logic above. util-linux
6446 upstream has been changed already to avoid this conflict,
6447 and we will re-add "-l" as soon as util-linux with this
6448 change has been released.
6450 * The dependency on libattr has been removed. Since a long
6451 time, the extended attribute calls have moved to glibc, and
6452 libattr is thus unnecessary.
6454 * Virtualization detection works without privileges now. This
6455 means the systemd-detect-virt binary no longer requires
6456 CAP_SYS_PTRACE file capabilities, and our daemons can run
6457 with fewer privileges.
6459 * systemd-networkd now runs under its own "systemd-network"
6460 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
6461 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
6462 loses the ability to write to files owned by root this way.
6464 * Similarly, systemd-resolved now runs under its own
6465 "systemd-resolve" user with no capabilities remaining.
6467 * Similarly, systemd-bus-proxyd now runs under its own
6468 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
6470 * systemd-networkd gained support for setting up "veth"
6471 virtual Ethernet devices for container connectivity, as well
6472 as GRE and VTI tunnels.
6474 * systemd-networkd will no longer automatically attempt to
6475 manually load kernel modules necessary for certain tunnel
6476 transports. Instead, it is assumed the kernel loads them
6477 automatically when required. This only works correctly on
6478 very new kernels. On older kernels, please consider adding
6479 the kernel modules to /etc/modules-load.d/ as a work-around.
6481 * The resolv.conf file systemd-resolved generates has been
6482 moved to /run/systemd/resolve/. If you have a symlink from
6483 /etc/resolv.conf, it might be necessary to correct it.
6485 * Two new service settings, ProtectHome= and ProtectSystem=,
6486 have been added. When enabled, they will make the user data
6487 (such as /home) inaccessible or read-only and the system
6488 (such as /usr) read-only, for specific services. This allows
6489 very light-weight per-service sandboxing to avoid
6490 modifications of user data or system files from
6491 services. These two new switches have been enabled for all
6492 of systemd's long-running services, where appropriate.
6494 * Socket units gained new SocketUser= and SocketGroup=
6495 settings to set the owner user and group of AF_UNIX sockets
6496 and FIFOs in the file system.
6498 * Socket units gained a new RemoveOnStop= setting. If enabled,
6499 all FIFOS and sockets in the file system will be removed
6500 when the specific socket unit is stopped.
6502 * Socket units gained a new Symlinks= setting. It takes a list
6503 of symlinks to create to file system sockets or FIFOs
6504 created by the specific Unix sockets. This is useful to
6505 manage symlinks to socket nodes with the same lifecycle as
6508 * The /dev/log socket and /dev/initctl FIFO have been moved to
6509 /run, and have been replaced by symlinks. This allows
6510 connecting to these facilities even if PrivateDevices=yes is
6511 used for a service (which makes /dev/log itself unavailable,
6512 but /run is left). This also has the benefit of ensuring
6513 that /dev only contains device nodes, directories and
6514 symlinks, and nothing else.
6516 * sd-daemon gained two new calls sd_pid_notify() and
6517 sd_pid_notifyf(). They are similar to sd_notify() and
6518 sd_notifyf(), but allow overriding of the source PID of
6519 notification messages if permissions permit this. This is
6520 useful to send notify messages on behalf of a different
6521 process (for example, the parent process). The
6522 systemd-notify tool has been updated to make use of this
6523 when sending messages (so that notification messages now
6524 originate from the shell script invoking systemd-notify and
6525 not the systemd-notify process itself. This should minimize
6526 a race where systemd fails to associate notification
6527 messages to services when the originating process already
6530 * A new "on-abnormal" setting for Restart= has been added. If
6531 set, it will result in automatic restarts on all "abnormal"
6532 reasons for a process to exit, which includes unclean
6533 signals, core dumps, timeouts and watchdog timeouts, but
6534 does not include clean and unclean exit codes or clean
6535 signals. Restart=on-abnormal is an alternative for
6536 Restart=on-failure for services that shall be able to
6537 terminate and avoid restarts on certain errors, by
6538 indicating so with an unclean exit code. Restart=on-failure
6539 or Restart=on-abnormal is now the recommended setting for
6540 all long-running services.
6542 * If the InaccessibleDirectories= service setting points to a
6543 mount point (or if there are any submounts contained within
6544 it), it is now attempted to completely unmount it, to make
6545 the file systems truly unavailable for the respective
6548 * The ReadOnlyDirectories= service setting and
6549 systemd-nspawn's --read-only parameter are now recursively
6550 applied to all submounts, too.
6552 * Mount units may now be created transiently via the bus APIs.
6554 * The support for SysV and LSB init scripts has been removed
6555 from the systemd daemon itself. Instead, it is now
6556 implemented as a generator that creates native systemd units
6557 from these scripts when needed. This enables us to remove a
6558 substantial amount of legacy code from PID 1, following the
6559 fact that many distributions only ship a very small number
6560 of LSB/SysV init scripts nowadays.
6562 * Privileged Xen (dom0) domains are not considered
6563 virtualization anymore by the virtualization detection
6564 logic. After all, they generally have unrestricted access to
6565 the hardware and usually are used to manage the unprivileged
6568 * systemd-tmpfiles gained a new "C" line type, for copying
6569 files or entire directories.
6571 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
6572 lines. So far, they have been non-globbing versions of the
6573 latter, and have thus been redundant. In future, it is
6574 recommended to only use "z". "m" has hence been removed
6575 from the documentation, even though it stays supported.
6577 * A tmpfiles snippet to recreate the most basic structure in
6578 /var has been added. This is enough to create the /var/run →
6579 /run symlink and create a couple of structural
6580 directories. This allows systems to boot up with an empty or
6581 volatile /var. Of course, while with this change, the core OS
6582 now is capable with dealing with a volatile /var, not all
6583 user services are ready for it. However, we hope that sooner
6584 or later, many service daemons will be changed upstream so
6585 that they are able to automatically create their necessary
6586 directories in /var at boot, should they be missing. This is
6587 the first step to allow state-less systems that only require
6588 the vendor image for /usr to boot.
6590 * systemd-nspawn has gained a new --tmpfs= switch to mount an
6591 empty tmpfs instance to a specific directory. This is
6592 particularly useful for making use of the automatic
6593 reconstruction of /var (see above), by passing --tmpfs=/var.
6595 * Access modes specified in tmpfiles snippets may now be
6596 prefixed with "~", which indicates that they shall be masked
6597 by whether the existing file or directory is currently
6598 writable, readable or executable at all. Also, if specified,
6599 the sgid/suid/sticky bits will be masked for all
6602 * A new passive target unit "network-pre.target" has been
6603 added which is useful for services that shall run before any
6604 network is configured, for example firewall scripts.
6606 * The "floppy" group that previously owned the /dev/fd*
6607 devices is no longer used. The "disk" group is now used
6608 instead. Distributions should probably deprecate usage of
6611 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
6612 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
6613 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
6614 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
6615 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
6616 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
6619 — Berlin, 2014-06-11
6623 * A new "systemd-timesyncd" daemon has been added for
6624 synchronizing the system clock across the network. It
6625 implements an SNTP client. In contrast to NTP
6626 implementations such as chrony or the NTP reference server,
6627 this only implements a client side, and does not bother with
6628 the full NTP complexity, focusing only on querying time from
6629 one remote server and synchronizing the local clock to
6630 it. Unless you intend to serve NTP to networked clients or
6631 want to connect to local hardware clocks, this simple NTP
6632 client should be more than appropriate for most
6633 installations. The daemon runs with minimal privileges, and
6634 has been hooked up with networkd to only operate when
6635 network connectivity is available. The daemon saves the
6636 current clock to disk every time a new NTP sync has been
6637 acquired, and uses this to possibly correct the system clock
6638 early at bootup, in order to accommodate for systems that
6639 lack an RTC such as the Raspberry Pi and embedded devices,
6640 and to make sure that time monotonically progresses on these
6641 systems, even if it is not always correct. To make use of
6642 this daemon, a new system user and group "systemd-timesync"
6643 needs to be created on installation of systemd.
6645 * The queue "seqnum" interface of libudev has been disabled, as
6646 it was generally incompatible with device namespacing as
6647 sequence numbers of devices go "missing" if the devices are
6648 part of a different namespace.
6650 * "systemctl list-timers" and "systemctl list-sockets" gained
6651 a --recursive switch for showing units of these types also
6652 for all local containers, similar in style to the already
6653 supported --recursive switch for "systemctl list-units".
6655 * A new RebootArgument= setting has been added for service
6656 units, which may be used to specify a kernel reboot argument
6657 to use when triggering reboots with StartLimitAction=.
6659 * A new FailureAction= setting has been added for service
6660 units which may be used to specify an operation to trigger
6661 when a service fails. This works similarly to
6662 StartLimitAction=, but unlike it, controls what is done
6663 immediately rather than only after several attempts to
6664 restart the service in question.
6666 * hostnamed got updated to also expose the kernel name,
6667 release, and version on the bus. This is useful for
6668 executing commands like hostnamectl with the -H switch.
6669 systemd-analyze makes use of this to properly display
6670 details when running non-locally.
6672 * The bootchart tool can now show cgroup information in the
6673 graphs it generates.
6675 * The CFS CPU quota cgroup attribute is now exposed for
6676 services. The new CPUQuota= switch has been added for this
6677 which takes a percentage value. Setting this will have the
6678 result that a service may never get more CPU time than the
6679 specified percentage, even if the machine is otherwise idle.
6681 * systemd-networkd learned IPIP and SIT tunnel support.
6683 * LSB init scripts exposing a dependency on $network will now
6684 get a dependency on network-online.target rather than simply
6685 network.target. This should bring LSB handling closer to
6686 what it was on SysV systems.
6688 * A new fsck.repair= kernel option has been added to control
6689 how fsck shall deal with unclean file systems at boot.
6691 * The (.ini) configuration file parser will now silently
6692 ignore sections whose name begins with "X-". This may be
6693 used to maintain application-specific extension sections in unit
6696 * machined gained a new API to query the IP addresses of
6697 registered containers. "machinectl status" has been updated
6698 to show these addresses in its output.
6700 * A new call sd_uid_get_display() has been added to the
6701 sd-login APIs for querying the "primary" session of a
6702 user. The "primary" session of the user is elected from the
6703 user's sessions and generally a graphical session is
6704 preferred over a text one.
6706 * A minimal systemd-resolved daemon has been added. It
6707 currently simply acts as a companion to systemd-networkd and
6708 manages resolv.conf based on per-interface DNS
6709 configuration, possibly supplied via DHCP. In the long run
6710 we hope to extend this into a local DNSSEC enabled DNS and
6713 * The systemd-networkd-wait-online tool is now enabled by
6714 default. It will delay network-online.target until a network
6715 connection has been configured. The tool primarily integrates
6716 with networkd, but will also make a best effort to make sense
6717 of network configuration performed in some other way.
6719 * Two new service options StartupCPUShares= and
6720 StartupBlockIOWeight= have been added that work similarly to
6721 CPUShares= and BlockIOWeight= however only apply during
6722 system startup. This is useful to prioritize certain services
6723 differently during bootup than during normal runtime.
6725 * hostnamed has been changed to prefer the statically
6726 configured hostname in /etc/hostname (unless set to
6727 'localhost' or empty) over any dynamic one supplied by
6728 dhcp. With this change, the rules for picking the hostname
6729 match more closely the rules of other configuration settings
6730 where the local administrator's configuration in /etc always
6731 overrides any other settings.
6733 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
6734 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
6735 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
6736 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
6737 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
6738 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
6739 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
6740 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
6741 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
6742 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
6743 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
6744 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
6745 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
6746 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
6747 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
6748 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6751 — Beijing, 2014-05-28
6755 * When restoring the screen brightness at boot, stay away from
6756 the darkest setting or from the lowest 5% of the available
6757 range, depending on which is the larger value of both. This
6758 should effectively protect the user from rebooting into a
6759 black screen, should the brightness have been set to minimum
6762 * sd-login gained a new sd_machine_get_class() call to
6763 determine the class ("vm" or "container") of a machine
6764 registered with machined.
6766 * sd-login gained new calls
6767 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
6768 to query the identity of the peer of a local AF_UNIX
6769 connection. They operate similarly to their sd_pid_get_xyz()
6772 * PID 1 will now maintain a system-wide system state engine
6773 with the states "starting", "running", "degraded",
6774 "maintenance", "stopping". These states are bound to system
6775 startup, normal runtime, runtime with at least one failed
6776 service, rescue/emergency mode and system shutdown. This
6777 state is shown in the "systemctl status" output when no unit
6778 name is passed. It is useful to determine system state, in
6779 particularly when doing so for many systems or containers at
6782 * A new command "list-machines" has been added to "systemctl"
6783 that lists all local OS containers and shows their system
6784 state (see above), if systemd runs inside of them.
6786 * systemctl gained a new "-r" switch to recursively enumerate
6787 units on all local containers, when used with the
6788 "list-unit" command (which is the default one that is
6789 executed when no parameters are specified).
6791 * The GPT automatic partition discovery logic will now honour
6792 two GPT partition flags: one may be set on a partition to
6793 cause it to be mounted read-only, and the other may be set
6794 on a partition to ignore it during automatic discovery.
6796 * Two new GPT type UUIDs have been added for automatic root
6797 partition discovery, for 32-bit and 64-bit ARM. This is not
6798 particularly useful for discovering the root directory on
6799 these architectures during bare-metal boots (since UEFI is
6800 not common there), but still very useful to allow booting of
6801 ARM disk images in nspawn with the -i option.
6803 * MAC addresses of interfaces created with nspawn's
6804 --network-interface= switch will now be generated from the
6805 machine name, and thus be stable between multiple invocations
6808 * logind will now automatically remove all IPC objects owned
6809 by a user if she or he fully logs out. This makes sure that
6810 users who are logged out cannot continue to consume IPC
6811 resources. This covers SysV memory, semaphores and message
6812 queues as well as POSIX shared memory and message
6813 queues. Traditionally, SysV and POSIX IPC had no lifecycle
6814 limits. With this functionality, that is corrected. This may
6815 be turned off by using the RemoveIPC= switch of logind.conf.
6817 * The systemd-machine-id-setup and tmpfiles tools gained a
6818 --root= switch to operate on a specific root directory,
6821 * journald can now forward logged messages to the TTYs of all
6822 logged in users ("wall"). This is the default for all
6823 emergency messages now.
6825 * A new tool systemd-journal-remote has been added to stream
6826 journal log messages across the network.
6828 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
6829 controller trees are mounted into it. Note that the
6830 directories mounted beneath it are not read-only. This is a
6831 security measure and is particularly useful because glibc
6832 actually includes a search logic to pick any tmpfs it can
6833 find to implement shm_open() if /dev/shm is not available
6834 (which it might very well be in namespaced setups).
6836 * machinectl gained a new "poweroff" command to cleanly power
6837 down a local OS container.
6839 * The PrivateDevices= unit file setting will now also drop the
6840 CAP_MKNOD capability from the capability bound set, and
6841 imply DevicePolicy=closed.
6843 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
6844 comprehensively on all long-running systemd services where
6845 this is appropriate.
6847 * systemd-udevd will now run in a disassociated mount
6848 namespace. To mount directories from udev rules, make sure to
6849 pull in mount units via SYSTEMD_WANTS properties.
6851 * The kdbus support gained support for uploading policy into
6852 the kernel. sd-bus gained support for creating "monitoring"
6853 connections that can eavesdrop into all bus communication
6854 for debugging purposes.
6856 * Timestamps may now be specified in seconds since the UNIX
6857 epoch Jan 1st, 1970 by specifying "@" followed by the value
6860 * Native tcpwrap support in systemd has been removed. tcpwrap
6861 is old code, not really maintained anymore and has serious
6862 shortcomings, and better options such as firewalls
6863 exist. For setups that require tcpwrap usage, please
6864 consider invoking your socket-activated service via tcpd,
6865 like on traditional inetd.
6867 * A new system.conf configuration option
6868 DefaultTimerAccuracySec= has been added that controls the
6869 default AccuracySec= setting of .timer units.
6871 * Timer units gained a new WakeSystem= switch. If enabled,
6872 timers configured this way will cause the system to resume
6873 from system suspend (if the system supports that, which most
6876 * Timer units gained a new Persistent= switch. If enabled,
6877 timers configured this way will save to disk when they have
6878 been last triggered. This information is then used on next
6879 reboot to possible execute overdue timer events, that
6880 could not take place because the system was powered off.
6881 This enables simple anacron-like behaviour for timer units.
6883 * systemctl's "list-timers" will now also list the time a
6884 timer unit was last triggered in addition to the next time
6885 it will be triggered.
6887 * systemd-networkd will now assign predictable IPv4LL
6888 addresses to its local interfaces.
6890 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
6891 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
6892 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
6893 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
6894 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
6895 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
6896 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
6897 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
6900 — Berlin, 2014-03-25
6904 * A new unit file setting RestrictAddressFamilies= has been
6905 added to restrict which socket address families unit
6906 processes gain access to. This takes address family names
6907 like "AF_INET" or "AF_UNIX", and is useful to minimize the
6908 attack surface of services via exotic protocol stacks. This
6909 is built on seccomp system call filters.
6911 * Two new unit file settings RuntimeDirectory= and
6912 RuntimeDirectoryMode= have been added that may be used to
6913 manage a per-daemon runtime directories below /run. This is
6914 an alternative for setting up directory permissions with
6915 tmpfiles snippets, and has the advantage that the runtime
6916 directory's lifetime is bound to the daemon runtime and that
6917 the daemon starts up with an empty directory each time. This
6918 is particularly useful when writing services that drop
6919 privileges using the User= or Group= setting.
6921 * The DeviceAllow= unit setting now supports globbing for
6922 matching against device group names.
6924 * The systemd configuration file system.conf gained new
6925 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
6926 DefaultMemoryAccounting= to globally turn on/off accounting
6927 for specific resources (cgroups) for all units. These
6928 settings may still be overridden individually in each unit
6931 * systemd-gpt-auto-generator is now able to discover /srv and
6932 root partitions in addition to /home and swap partitions. It
6933 also supports LUKS-encrypted partitions now. With this in
6934 place, automatic discovery of partitions to mount following
6935 the Discoverable Partitions Specification
6936 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
6937 is now a lot more complete. This allows booting without
6938 /etc/fstab and without root= on the kernel command line on
6939 systems prepared appropriately.
6941 * systemd-nspawn gained a new --image= switch which allows
6942 booting up disk images and Linux installations on any block
6943 device that follow the Discoverable Partitions Specification
6944 (see above). This means that installations made with
6945 appropriately updated installers may now be started and
6946 deployed using container managers, completely
6947 unmodified. (We hope that libvirt-lxc will add support for
6948 this feature soon, too.)
6950 * systemd-nspawn gained a new --network-macvlan= setting to
6951 set up a private macvlan interface for the
6952 container. Similarly, systemd-networkd gained a new
6953 Kind=macvlan setting in .netdev files.
6955 * systemd-networkd now supports configuring local addresses
6958 * A new tool systemd-network-wait-online has been added to
6959 synchronously wait for network connectivity using
6962 * The sd-bus.h bus API gained a new sd_bus_track object for
6963 tracking the lifecycle of bus peers. Note that sd-bus.h is
6964 still not a public API though (unless you specify
6965 --enable-kdbus on the configure command line, which however
6966 voids your warranty and you get no API stability guarantee).
6968 * The $XDG_RUNTIME_DIR runtime directories for each user are
6969 now individual tmpfs instances, which has the benefit of
6970 introducing separate pools for each user, with individual
6971 size limits, and thus making sure that unprivileged clients
6972 can no longer negatively impact the system or other users by
6973 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
6974 RuntimeDirectorySize= has been introduced that allows
6975 controlling the default size limit for all users. It
6976 defaults to 10% of the available physical memory. This is no
6977 replacement for quotas on tmpfs though (which the kernel
6978 still does not support), as /dev/shm and /tmp are still
6979 shared resources used by both the system and unprivileged
6982 * logind will now automatically turn off automatic suspending
6983 on laptop lid close when more than one display is
6984 connected. This was previously expected to be implemented
6985 individually in desktop environments (such as GNOME),
6986 however has been added to logind now, in order to fix a
6987 boot-time race where a desktop environment might not have
6988 been started yet and thus not been able to take an inhibitor
6989 lock at the time where logind already suspends the system
6990 due to a closed lid.
6992 * logind will now wait at least 30s after each system
6993 suspend/resume cycle, and 3min after system boot before
6994 suspending the system due to a closed laptop lid. This
6995 should give USB docking stations and similar enough time to
6996 be probed and configured after system resume and boot in
6997 order to then act as suspend blocker.
6999 * systemd-run gained a new --property= setting which allows
7000 initialization of resource control properties (and others)
7001 for the created scope or service unit. Example: "systemd-run
7002 --property=BlockIOWeight=10 updatedb" may be used to run
7003 updatedb at a low block IO scheduling weight.
7005 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
7006 now also work in --scope mode.
7008 * When systemd is compiled with kdbus support, basic support
7009 for enforced policies is now in place. (Note that enabling
7010 kdbus still voids your warranty and no API compatibility
7013 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
7014 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7015 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
7016 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
7017 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
7018 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
7019 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
7020 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
7021 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
7022 Zbigniew Jędrzejewski-Szmek
7024 — Berlin, 2014-03-12
7028 * systemd will now relabel /dev after loading the SMACK policy
7029 according to SMACK rules.
7031 * A new unit file option AppArmorProfile= has been added to
7032 set the AppArmor profile for the processes of a unit.
7034 * A new condition check ConditionArchitecture= has been added
7035 to conditionalize units based on the system architecture, as
7036 reported by uname()'s "machine" field.
7038 * systemd-networkd now supports matching on the system
7039 virtualization, architecture, kernel command line, host name
7042 * logind is now a lot more aggressive when suspending the
7043 machine due to a closed laptop lid. Instead of acting only
7044 on the lid close action, it will continuously watch the lid
7045 status and act on it. This is useful for laptops where the
7046 power button is on the outside of the chassis so that it can
7047 be reached without opening the lid (such as the Lenovo
7048 Yoga). On those machines, logind will now immediately
7049 re-suspend the machine if the power button has been
7050 accidentally pressed while the laptop was suspended and in a
7051 backpack or similar.
7053 * logind will now watch SW_DOCK switches and inhibit reaction
7054 to the lid switch if it is pressed. This means that logind
7055 will not suspend the machine anymore if the lid is closed
7056 and the system is docked, if the laptop supports SW_DOCK
7057 notifications via the input layer. Note that ACPI docking
7058 stations do not generate this currently. Also note that this
7059 logic is usually not fully sufficient and Desktop
7060 Environments should take a lid switch inhibitor lock when an
7061 external display is connected, as systemd will not watch
7064 * nspawn will now make use of the devices cgroup controller by
7065 default, and only permit creation of and access to the usual
7066 API device nodes like /dev/null or /dev/random, as well as
7067 access to (but not creation of) the pty devices.
7069 * We will now ship a default .network file for
7070 systemd-networkd that automatically configures DHCP for
7071 network interfaces created by nspawn's --network-veth or
7072 --network-bridge= switches.
7074 * systemd will now understand the usual M, K, G, T suffixes
7075 according to SI conventions (i.e. to the base 1000) when
7076 referring to throughput and hardware metrics. It will stay
7077 with IEC conventions (i.e. to the base 1024) for software
7078 metrics, according to what is customary according to
7079 Wikipedia. We explicitly document which base applies for
7080 each configuration option.
7082 * The DeviceAllow= setting in unit files now supports a syntax
7083 to whitelist an entire group of devices node majors at once,
7084 based on the /proc/devices listing. For example, with the
7085 string "char-pts", it is now possible to whitelist all
7086 current and future pseudo-TTYs at once.
7088 * sd-event learned a new "post" event source. Event sources of
7089 this type are triggered by the dispatching of any event
7090 source of a type that is not "post". This is useful for
7091 implementing clean-up and check event sources that are
7092 triggered by other work being done in the program.
7094 * systemd-networkd is no longer statically enabled, but uses
7095 the usual [Install] sections so that it can be
7096 enabled/disabled using systemctl. It still is enabled by
7099 * When creating a veth interface pair with systemd-nspawn, the
7100 host side will now be prefixed with "vb-" if
7101 --network-bridge= is used, and with "ve-" if --network-veth
7102 is used. This way, it is easy to distinguish these cases on
7103 the host, for example to apply different configuration to
7104 them with systemd-networkd.
7106 * The compatibility libraries for libsystemd-journal.so,
7107 libsystem-id128.so, libsystemd-login.so and
7108 libsystemd-daemon.so do not make use of IFUNC
7109 anymore. Instead, we now build libsystemd.so multiple times
7110 under these alternative names. This means that the footprint
7111 is drastically increased, but given that these are
7112 transitional compatibility libraries, this should not matter
7113 much. This change has been made necessary to support the ARM
7114 platform for these compatibility libraries, as the ARM
7115 toolchain is not really at the same level as the toolchain
7116 for other architectures like x86 and does not support
7117 IFUNC. Please make sure to use --enable-compat-libs only
7118 during a transitional period!
7120 * The .include syntax has been deprecated and is not documented
7121 anymore. Drop-in files in .d directories should be used instead.
7123 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
7124 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7125 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
7126 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
7127 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
7128 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
7129 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
7130 Zbigniew Jędrzejewski-Szmek
7132 — Berlin, 2014-02-24
7136 * A new component "systemd-networkd" has been added that can
7137 be used to configure local network interfaces statically or
7138 via DHCP. It is capable of bringing up bridges, VLANs, and
7139 bonding. Currently, no hook-ups for interactive network
7140 configuration are provided. Use this for your initrd,
7141 container, embedded, or server setup if you need a simple,
7142 yet powerful, network configuration solution. This
7143 configuration subsystem is quite nifty, as it allows wildcard
7144 hotplug matching in interfaces. For example, with a single
7145 configuration snippet, you can configure that all Ethernet
7146 interfaces showing up are automatically added to a bridge,
7147 or similar. It supports link-sensing and more.
7149 * A new tool "systemd-socket-proxyd" has been added which can
7150 act as a bidirectional proxy for TCP sockets. This is
7151 useful for adding socket activation support to services that
7152 do not actually support socket activation, including virtual
7153 machines and the like.
7155 * Add a new tool to save/restore rfkill state on
7158 * Save/restore state of keyboard backlights in addition to
7159 display backlights on shutdown/boot.
7161 * udev learned a new SECLABEL{} construct to label device
7162 nodes with a specific security label when they appear. For
7163 now, only SECLABEL{selinux} is supported, but the syntax is
7164 prepared for additional security frameworks.
7166 * udev gained a new scheme to configure link-level attributes
7167 from files in /etc/systemd/network/*.link. These files can
7168 match against MAC address, device path, driver name and type,
7169 and will apply attributes like the naming policy, link speed,
7170 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
7171 address assignment policy (randomized, ...).
7173 * The configuration of network interface naming rules for
7174 "permanent interface names" has changed: a new NamePolicy=
7175 setting in the [Link] section of .link files determines the
7176 priority of possible naming schemes (onboard, slot, MAC,
7177 path). The default value of this setting is determined by
7178 /usr/lib/net/links/99-default.link. Old
7179 80-net-name-slot.rules udev configuration file has been
7180 removed, so local configuration overriding this file should
7181 be adapted to override 99-default.link instead.
7183 * When the User= switch is used in a unit file, also
7184 initialize $SHELL= based on the user database entry.
7186 * systemd no longer depends on libdbus. All communication is
7187 now done with sd-bus, systemd's low-level bus library
7190 * kdbus support has been added to PID 1 itself. When kdbus is
7191 enabled, this causes PID 1 to set up the system bus and
7192 enable support for a new ".busname" unit type that
7193 encapsulates bus name activation on kdbus. It works a little
7194 bit like ".socket" units, except for bus names. A new
7195 generator has been added that converts classic dbus1 service
7196 activation files automatically into native systemd .busname
7199 * sd-bus: add a light-weight vtable implementation that allows
7200 defining objects on the bus with a simple static const
7201 vtable array of its methods, signals and properties.
7203 * systemd will not generate or install static dbus
7204 introspection data anymore to /usr/share/dbus-1/interfaces,
7205 as the precise format of these files is unclear, and
7206 nothing makes use of it.
7208 * A proxy daemon is now provided to proxy clients connecting
7209 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
7210 compatibility with classic D-Bus.
7212 * A bus driver implementation has been added that supports the
7213 classic D-Bus bus driver calls on kdbus, also for
7214 compatibility purposes.
7216 * A new API "sd-event.h" has been added that implements a
7217 minimal event loop API built around epoll. It provides a
7218 couple of features that direct epoll usage is lacking:
7219 prioritization of events, scales to large numbers of timer
7220 events, per-event timer slack (accuracy), system-wide
7221 coalescing of timer events, exit handlers, watchdog
7222 supervision support using systemd's sd_notify() API, child
7225 * A new API "sd-rntl.h" has been added that provides an API
7226 around the route netlink interface of the kernel, similar in
7227 style to "sd-bus.h".
7229 * A new API "sd-dhcp-client.h" has been added that provides a
7230 small DHCPv4 client-side implementation. This is used by
7233 * There is a new kernel command line option
7234 "systemd.restore_state=0|1". When set to "0", none of the
7235 systemd tools will restore saved runtime state to hardware
7236 devices. More specifically, the rfkill and backlight states
7239 * The FsckPassNo= compatibility option in mount/service units
7240 has been removed. The fstab generator will now add the
7241 necessary dependencies automatically, and does not require
7242 PID1's support for that anymore.
7244 * journalctl gained a new switch, --list-boots, that lists
7245 recent boots with their times and boot IDs.
7247 * The various tools like systemctl, loginctl, timedatectl,
7248 busctl, systemd-run, ... have gained a new switch "-M" to
7249 connect to a specific, local OS container (as direct
7250 connection, without requiring SSH). This works on any
7251 container that is registered with machined, such as those
7252 created by libvirt-lxc or nspawn.
7254 * systemd-run and systemd-analyze also gained support for "-H"
7255 to connect to remote hosts via SSH. This is particularly
7256 useful for systemd-run because it enables queuing of jobs
7257 onto remote systems.
7259 * machinectl gained a new command "login" to open a getty
7260 login in any local container. This works with any container
7261 that is registered with machined (such as those created by
7262 libvirt-lxc or nspawn), and which runs systemd inside.
7264 * machinectl gained a new "reboot" command that may be used to
7265 trigger a reboot on a specific container that is registered
7266 with machined. This works on any container that runs an init
7267 system of some kind.
7269 * systemctl gained a new "list-timers" command to print a nice
7270 listing of installed timer units with the times they elapse
7273 * Alternative reboot() parameters may now be specified on the
7274 "systemctl reboot" command line and are passed to the
7275 reboot() system call.
7277 * systemctl gained a new --job-mode= switch to configure the
7278 mode to queue a job with. This is a more generic version of
7279 --fail, --irreversible, and --ignore-dependencies, which are
7280 still available but not advertised anymore.
7282 * /etc/systemd/system.conf gained new settings to configure
7283 various default timeouts of units, as well as the default
7284 start limit interval and burst. These may still be overridden
7287 * PID1 will now export on the bus profile data of the security
7288 policy upload process (such as the SELinux policy upload to
7291 * journald: when forwarding logs to the console, include
7292 timestamps (following the setting in
7293 /sys/module/printk/parameters/time).
7295 * OnCalendar= in timer units now understands the special
7296 strings "yearly" and "annually". (Both are equivalent)
7298 * The accuracy of timer units is now configurable with the new
7299 AccuracySec= setting. It defaults to 1min.
7301 * A new dependency type JoinsNamespaceOf= has been added that
7302 allows running two services within the same /tmp and network
7303 namespace, if PrivateNetwork= or PrivateTmp= are used.
7305 * A new command "cat" has been added to systemctl. It outputs
7306 the original unit file of a unit, and concatenates the
7307 contents of additional "drop-in" unit file snippets, so that
7308 the full configuration is shown.
7310 * systemctl now supports globbing on the various "list-xyz"
7311 commands, like "list-units" or "list-sockets", as well as on
7312 those commands which take multiple unit names.
7314 * journalctl's --unit= switch gained support for globbing.
7316 * All systemd daemons now make use of the watchdog logic so
7317 that systemd automatically notices when they hang.
7319 * If the $container_ttys environment variable is set,
7320 getty-generator will automatically spawn a getty for each
7321 listed tty. This is useful for container managers to request
7322 login gettys to be spawned on as many ttys as needed.
7324 * %h, %s, %U specifier support is not available anymore when
7325 used in unit files for PID 1. This is because NSS calls are
7326 not safe from PID 1. They stay available for --user
7327 instances of systemd, and as special case for the root user.
7329 * loginctl gained a new "--no-legend" switch to turn off output
7332 * The "sd-login.h" API gained three new calls:
7333 sd_session_is_remote(), sd_session_get_remote_user(),
7334 sd_session_get_remote_host() to query information about
7337 * The udev hardware database now also carries vendor/product
7338 information of SDIO devices.
7340 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
7341 determine whether watchdog notifications are requested by
7344 * Socket-activated per-connection services now include a
7345 short description of the connection parameters in the
7348 * tmpfiles gained a new "--boot" option. When this is not used,
7349 only lines where the command character is not suffixed with
7350 "!" are executed. When this option is specified, those
7351 options are executed too. This partitions tmpfiles
7352 directives into those that can be safely executed at any
7353 time, and those which should be run only at boot (for
7354 example, a line that creates /run/nologin).
7356 * A new API "sd-resolve.h" has been added which provides a simple
7357 asynchronous wrapper around glibc NSS host name resolution
7358 calls, such as getaddrinfo(). In contrast to glibc's
7359 getaddrinfo_a(), it does not use signals. In contrast to most
7360 other asynchronous name resolution libraries, this one does
7361 not reimplement DNS, but reuses NSS, so that alternate
7362 host name resolution systems continue to work, such as mDNS,
7363 LDAP, etc. This API is based on libasyncns, but it has been
7364 cleaned up for inclusion in systemd.
7366 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
7367 "sd-daemon.h" are no longer found in individual libraries
7368 libsystemd-journal.so, libsystemd-login.so,
7369 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
7370 merged them into a single library, libsystemd.so, which
7371 provides all symbols. The reason for this is cyclic
7372 dependencies, as these libraries tend to use each other's
7373 symbols. So far, we have managed to workaround that by linking
7374 a copy of a good part of our code into each of these
7375 libraries again and again, which, however, makes certain
7376 things hard to do, like sharing static variables. Also, it
7377 substantially increases footprint. With this change, there
7378 is only one library for the basic APIs systemd
7379 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
7380 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
7381 library as well, however are subject to the --enable-kdbus
7382 switch (see below). Note that "sd-dhcp-client.h" is not part
7383 of this library (this is because it only consumes, never
7384 provides, services of/to other APIs). To make the transition
7385 easy from the separate libraries to the unified one, we
7386 provide the --enable-compat-libs compile-time switch which
7387 will generate stub libraries that are compatible with the
7388 old ones but redirect all calls to the new one.
7390 * All of the kdbus logic and the new APIs "sd-bus.h",
7391 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
7392 and "sd-utf8.h" are compile-time optional via the
7393 "--enable-kdbus" switch, and they are not compiled in by
7394 default. To make use of kdbus, you have to explicitly enable
7395 the switch. Note however, that neither the kernel nor the
7396 userspace API for all of this is considered stable yet. We
7397 want to maintain the freedom to still change the APIs for
7398 now. By specifying this build-time switch, you acknowledge
7399 that you are aware of the instability of the current
7402 * Also, note that while kdbus is pretty much complete,
7403 it lacks one thing: proper policy support. This means you
7404 can build a fully working system with all features; however,
7405 it will be highly insecure. Policy support will be added in
7406 one of the next releases, at the same time that we will
7407 declare the APIs stable.
7409 * When the kernel command line argument "kdbus" is specified,
7410 systemd will automatically load the kdbus.ko kernel module. At
7411 this stage of development, it is only useful for testing kdbus
7412 and should not be used in production. Note: if "--enable-kdbus"
7413 is specified, and the kdbus.ko kernel module is available, and
7414 "kdbus" is added to the kernel command line, the entire system
7415 runs with kdbus instead of dbus-daemon, with the above mentioned
7416 problem of missing the system policy enforcement. Also a future
7417 version of kdbus.ko or a newer systemd will not be compatible with
7418 each other, and will unlikely be able to boot the machine if only
7419 one of them is updated.
7421 * systemctl gained a new "import-environment" command which
7422 uploads the caller's environment (or parts thereof) into the
7423 service manager so that it is inherited by services started
7424 by the manager. This is useful to upload variables like
7425 $DISPLAY into the user service manager.
7427 * A new PrivateDevices= switch has been added to service units
7428 which allows running a service with a namespaced /dev
7429 directory that does not contain any device nodes for
7430 physical devices. More specifically, it only includes devices
7431 such as /dev/null, /dev/urandom, and /dev/zero which are API
7434 * logind has been extended to support behaviour like VT
7435 switching on seats that do not support a VT. This makes
7436 multi-session available on seats that are not the first seat
7437 (seat0), and on systems where kernel support for VTs has
7438 been disabled at compile-time.
7440 * If a process holds a delay lock for system sleep or shutdown
7441 and fails to release it in time, we will now log its
7442 identity. This makes it easier to identify processes that
7443 cause slow suspends or power-offs.
7445 * When parsing /etc/crypttab, support for a new key-slot=
7446 option as supported by Debian is added. It allows indicating
7447 which LUKS slot to use on disk, speeding up key loading.
7449 * The sd_journal_sendv() API call has been checked and
7450 officially declared to be async-signal-safe so that it may
7451 be invoked from signal handlers for logging purposes.
7453 * Boot-time status output is now enabled automatically after a
7454 short timeout if boot does not progress, in order to give
7455 the user an indication what she or he is waiting for.
7457 * The boot-time output has been improved to show how much time
7458 remains until jobs expire.
7460 * The KillMode= switch in service units gained a new possible
7461 value "mixed". If set, and the unit is shut down, then the
7462 initial SIGTERM signal is sent only to the main daemon
7463 process, while the following SIGKILL signal is sent to
7464 all remaining processes of the service.
7466 * When a scope unit is registered, a new property "Controller"
7467 may be set. If set to a valid bus name, systemd will send a
7468 RequestStop() signal to this name when it would like to shut
7469 down the scope. This may be used to hook manager logic into
7470 the shutdown logic of scope units. Also, scope units may now
7471 be put in a special "abandoned" state, in which case the
7472 manager process which created them takes no further
7473 responsibilities for it.
7475 * When reading unit files, systemd will now verify
7476 the access mode of these files, and warn about certain
7477 suspicious combinations. This has been added to make it
7478 easier to track down packaging bugs where unit files are
7479 marked executable or world-writable.
7481 * systemd-nspawn gained a new "--setenv=" switch to set
7482 container-wide environment variables. The similar option in
7483 systemd-activate was renamed from "--environment=" to
7484 "--setenv=" for consistency.
7486 * systemd-nspawn has been updated to create a new kdbus domain
7487 for each container that is invoked, thus allowing each
7488 container to have its own set of system and user buses,
7489 independent of the host.
7491 * systemd-nspawn gained a new --drop-capability= switch to run
7492 the container with less capabilities than the default. Both
7493 --drop-capability= and --capability= now take the special
7494 string "all" for dropping or keeping all capabilities.
7496 * systemd-nspawn gained new switches for executing containers
7497 with specific SELinux labels set.
7499 * systemd-nspawn gained a new --quiet switch to not generate
7500 any additional output but the container's own console
7503 * systemd-nspawn gained a new --share-system switch to run a
7504 container without PID namespacing enabled.
7506 * systemd-nspawn gained a new --register= switch to control
7507 whether the container is registered with systemd-machined or
7508 not. This is useful for containers that do not run full
7509 OS images, but only specific apps.
7511 * systemd-nspawn gained a new --keep-unit which may be used
7512 when invoked as the only program from a service unit, and
7513 results in registration of the unit service itself in
7514 systemd-machined, instead of a newly opened scope unit.
7516 * systemd-nspawn gained a new --network-interface= switch for
7517 moving arbitrary interfaces to the container. The new
7518 --network-veth switch creates a virtual Ethernet connection
7519 between host and container. The new --network-bridge=
7520 switch then allows assigning the host side of this virtual
7521 Ethernet connection to a bridge device.
7523 * systemd-nspawn gained a new --personality= switch for
7524 setting the kernel personality for the container. This is
7525 useful when running a 32-bit container on a 64-bit host. A
7526 similar option Personality= is now also available for service
7529 * logind will now also track a "Desktop" identifier for each
7530 session which encodes the desktop environment of it. This is
7531 useful for desktop environments that want to identify
7532 multiple running sessions of itself easily.
7534 * A new SELinuxContext= setting for service units has been
7535 added that allows setting a specific SELinux execution
7536 context for a service.
7538 * Most systemd client tools will now honour $SYSTEMD_LESS for
7539 settings of the "less" pager. By default, these tools will
7540 override $LESS to allow certain operations to work, such as
7541 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
7542 influence this logic.
7544 * systemd's "seccomp" hook-up has been changed to make use of
7545 the libseccomp library instead of using its own
7546 implementation. This has benefits for portability among
7549 * For usage together with SystemCallFilter=, a new
7550 SystemCallErrorNumber= setting has been introduced that
7551 allows configuration of a system error number to be returned
7552 on filtered system calls, instead of immediately killing the
7553 process. Also, SystemCallArchitectures= has been added to
7554 limit access to system calls of a particular architecture
7555 (in order to turn off support for unused secondary
7556 architectures). There is also a global
7557 SystemCallArchitectures= setting in system.conf now to turn
7558 off support for non-native system calls system-wide.
7560 * systemd requires a kernel with a working name_to_handle_at(),
7561 please see the kernel config requirements in the README file.
7563 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
7564 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
7565 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
7566 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
7567 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
7568 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
7569 Elia Pinto, Florian Weimer, George McCollister, Goffredo
7570 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
7571 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
7572 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
7573 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
7574 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
7575 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
7576 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
7577 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
7578 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
7579 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
7580 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
7581 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
7582 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
7583 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
7584 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
7585 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
7586 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
7588 — Berlin, 2014-02-20
7592 * logind has gained support for facilitating privileged input
7593 and drm device access for unprivileged clients. This work is
7594 useful to allow Wayland display servers (and similar
7595 programs, such as kmscon) to run under the user's ID and
7596 access input and drm devices which are normally
7597 protected. When this is used (and the kernel is new enough)
7598 logind will "mute" IO on the file descriptors passed to
7599 Wayland as long as it is in the background and "unmute" it
7600 if it returns into the foreground. This allows secure
7601 session switching without allowing background sessions to
7602 eavesdrop on input and display data. This also introduces
7603 session switching support if VT support is turned off in the
7604 kernel, and on seats that are not seat0.
7606 * A new kernel command line option luks.options= is understood
7607 now which allows specifying LUKS options for usage for LUKS
7608 encrypted partitions specified with luks.uuid=.
7610 * tmpfiles.d(5) snippets may now use specifier expansion in
7611 path names. More specifically %m, %b, %H, %v, are now
7612 replaced by the local machine id, boot id, hostname, and
7613 kernel version number.
7615 * A new tmpfiles.d(5) command "m" has been introduced which
7616 may be used to change the owner/group/access mode of a file
7617 or directory if it exists, but do nothing if it does not.
7619 * This release removes high-level support for the
7620 MemorySoftLimit= cgroup setting. The underlying kernel
7621 cgroup attribute memory.soft_limit= is currently badly
7622 designed and likely to be removed from the kernel API in its
7623 current form, hence we should not expose it for now.
7625 * The memory.use_hierarchy cgroup attribute is now enabled for
7626 all cgroups systemd creates in the memory cgroup
7627 hierarchy. This option is likely to be come the built-in
7628 default in the kernel anyway, and the non-hierarchical mode
7629 never made much sense in the intrinsically hierarchical
7632 * A new field _SYSTEMD_SLICE= is logged along with all journal
7633 messages containing the slice a message was generated
7634 from. This is useful to allow easy per-customer filtering of
7635 logs among other things.
7637 * systemd-journald will no longer adjust the group of journal
7638 files it creates to the "systemd-journal" group. Instead we
7639 rely on the journal directory to be owned by the
7640 "systemd-journal" group, and its setgid bit set, so that the
7641 kernel file system layer will automatically enforce that
7642 journal files inherit this group assignment. The reason for
7643 this change is that we cannot allow NSS look-ups from
7644 journald which would be necessary to resolve
7645 "systemd-journal" to a numeric GID, because this might
7646 create deadlocks if NSS involves synchronous queries to
7647 other daemons (such as nscd, or sssd) which in turn are
7648 logging clients of journald and might block on it, which
7649 would then dead lock. A tmpfiles.d(5) snippet included in
7650 systemd will make sure the setgid bit and group are
7651 properly set on the journal directory if it exists on every
7652 boot. However, we recommend adjusting it manually after
7653 upgrades too (or from RPM scriptlets), so that the change is
7654 not delayed until next reboot.
7656 * Backlight and random seed files in /var/lib/ have moved into
7657 the /var/lib/systemd/ directory, in order to centralize all
7658 systemd generated files in one directory.
7660 * Boot time performance measurements (as displayed by
7661 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
7662 performance information if that's available to determine how
7663 much time BIOS and boot loader initialization required. With
7664 a sufficiently new BIOS you hence no longer need to boot
7665 with Gummiboot to get access to such information.
7667 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
7668 Cristian Rodríguez, Dave Reisner, David Herrmann, David
7669 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
7670 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
7671 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
7672 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
7673 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7675 — Berlin, 2013-10-02
7679 * The Restart= option for services now understands a new
7680 on-watchdog setting, which will restart the service
7681 automatically if the service stops sending out watchdog keep
7682 alive messages (as configured with WatchdogSec=).
7684 * The getty generator (which is responsible for bringing up a
7685 getty on configured serial consoles) will no longer only
7686 start a getty on the primary kernel console but on all
7687 others, too. This makes the order in which console= is
7688 specified on the kernel command line less important.
7690 * libsystemd-logind gained a new sd_session_get_vt() call to
7691 retrieve the VT number of a session.
7693 * If the option "tries=0" is set for an entry of /etc/crypttab
7694 its passphrase is queried indefinitely instead of any
7695 maximum number of tries.
7697 * If a service with a configure PID file terminates its PID
7698 file will now be removed automatically if it still exists
7699 afterwards. This should put an end to stale PID files.
7701 * systemd-run will now also take relative binary path names
7702 for execution and no longer insists on absolute paths.
7704 * InaccessibleDirectories= and ReadOnlyDirectories= now take
7705 paths that are optionally prefixed with "-" to indicate that
7706 it should not be considered a failure if they do not exist.
7708 * journalctl -o (and similar commands) now understands a new
7709 output mode "short-precise", it is similar to "short" but
7710 shows timestamps with usec accuracy.
7712 * The option "discard" (as known from Debian) is now
7713 synonymous to "allow-discards" in /etc/crypttab. In fact,
7714 "discard" is preferred now (since it is easier to remember
7717 * Some licensing clean-ups were made, so that more code is now
7718 LGPL-2.1 licensed than before.
7720 * A minimal tool to save/restore the display backlight
7721 brightness across reboots has been added. It will store the
7722 backlight setting as late as possible at shutdown, and
7723 restore it as early as possible during reboot.
7725 * A logic to automatically discover and enable home and swap
7726 partitions on GPT disks has been added. With this in place
7727 /etc/fstab becomes optional for many setups as systemd can
7728 discover certain partitions located on the root disk
7729 automatically. Home partitions are recognized under their
7730 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
7731 partitions are recognized under their GPT type ID
7732 0657fd6da4ab43c484e50933c84b4f4f.
7734 * systemd will no longer pass any environment from the kernel
7735 or initrd to system services. If you want to set an
7736 environment for all services, do so via the kernel command
7737 line systemd.setenv= assignment.
7739 * The systemd-sysctl tool no longer natively reads the file
7740 /etc/sysctl.conf. If desired, the file should be symlinked
7741 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
7742 legacy support by a symlink rather than built-in code, it
7743 also makes the otherwise hidden order of application of the
7744 different files visible. (Note that this partly reverts to a
7745 pre-198 application order of sysctl knobs!)
7747 * The "systemctl set-log-level" and "systemctl dump" commands
7748 have been moved to systemd-analyze.
7750 * systemd-run learned the new --remain-after-exit switch,
7751 which causes the scope unit not to be cleaned up
7752 automatically after the process terminated.
7754 * tmpfiles learned a new --exclude-prefix= switch to exclude
7755 certain paths from operation.
7757 * journald will now automatically flush all messages to disk
7758 as soon as a message at the log level CRIT, ALERT or EMERG
7761 Contributions from: Andrew Cook, Brandon Philips, Christian
7762 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
7763 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
7764 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
7765 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
7766 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
7767 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
7768 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
7769 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
7770 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
7771 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
7772 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
7773 William Giokas, Zbigniew Jędrzejewski-Szmek
7775 — Berlin, 2013-09-13
7779 * The documentation has been updated to cover the various new
7780 concepts introduced with 205.
7782 * Unit files now understand the new %v specifier which
7783 resolves to the kernel version string as returned by "uname
7786 * systemctl now supports filtering the unit list output by
7787 load state, active state and sub state, using the new
7790 * "systemctl status" will now show the results of the
7791 condition checks (like ConditionPathExists= and similar) of
7792 the last start attempts of the unit. They are also logged to
7795 * "journalctl -b" may now be used to look for boot output of a
7796 specific boot. Try "journalctl -b -1" for the previous boot,
7797 but the syntax is substantially more powerful.
7799 * "journalctl --show-cursor" has been added which prints the
7800 cursor string the last shown log line. This may then be used
7801 with the new "journalctl --after-cursor=" switch to continue
7802 browsing logs from that point on.
7804 * "journalctl --force" may now be used to force regeneration
7807 * Creation of "dead" device nodes has been moved from udev
7808 into kmod and tmpfiles. Previously, udev would read the kmod
7809 databases to pre-generate dead device nodes based on meta
7810 information contained in kernel modules, so that these would
7811 be auto-loaded on access rather then at boot. As this
7812 does not really have much to do with the exposing actual
7813 kernel devices to userspace this has always been slightly
7814 alien in the udev codebase. Following the new scheme kmod
7815 will now generate a runtime snippet for tmpfiles from the
7816 module meta information and it now is tmpfiles' job to the
7817 create the nodes. This also allows overriding access and
7818 other parameters for the nodes using the usual tmpfiles
7819 facilities. As side effect this allows us to remove the
7820 CAP_SYS_MKNOD capability bit from udevd entirely.
7822 * logind's device ACLs may now be applied to these "dead"
7823 devices nodes too, thus finally allowing managed access to
7824 devices such as /dev/snd/sequencer without loading the
7825 backing module right-away.
7827 * A new RPM macro has been added that may be used to apply
7828 tmpfiles configuration during package installation.
7830 * systemd-detect-virt and ConditionVirtualization= now can
7831 detect User-Mode-Linux machines (UML).
7833 * journald will now implicitly log the effective capabilities
7834 set of processes in the message metadata.
7836 * systemd-cryptsetup has gained support for TrueCrypt volumes.
7838 * The initrd interface has been simplified (more specifically,
7839 support for passing performance data via environment
7840 variables and fsck results via files in /run has been
7841 removed). These features were non-essential, and are
7842 nowadays available in a much nicer way by having systemd in
7843 the initrd serialize its state and have the hosts systemd
7844 deserialize it again.
7846 * The udev "keymap" data files and tools to apply keyboard
7847 specific mappings of scan to key codes, and force-release
7848 scan code lists have been entirely replaced by a udev
7849 "keyboard" builtin and a hwdb data file.
7851 * systemd will now honour the kernel's "quiet" command line
7852 argument also during late shutdown, resulting in a
7853 completely silent shutdown when used.
7855 * There's now an option to control the SO_REUSEPORT socket
7856 option in .socket units.
7858 * Instance units will now automatically get a per-template
7859 subslice of system.slice unless something else is explicitly
7860 configured. For example, instances of sshd@.service will now
7861 implicitly be placed in system-sshd.slice rather than
7862 system.slice as before.
7864 * Test coverage support may now be enabled at build time.
7866 Contributions from: Dave Reisner, Frederic Crozat, Harald
7867 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
7868 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
7869 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
7870 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
7871 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
7872 Giokas, Zbigniew Jędrzejewski-Szmek
7874 — Berlin, 2013-07-23
7878 * Two new unit types have been introduced:
7880 Scope units are very similar to service units, however, are
7881 created out of pre-existing processes — instead of PID 1
7882 forking off the processes. By using scope units it is
7883 possible for system services and applications to group their
7884 own child processes (worker processes) in a powerful way
7885 which then maybe used to organize them, or kill them
7886 together, or apply resource limits on them.
7888 Slice units may be used to partition system resources in an
7889 hierarchical fashion and then assign other units to them. By
7890 default there are now three slices: system.slice (for all
7891 system services), user.slice (for all user sessions),
7892 machine.slice (for VMs and containers).
7894 Slices and scopes have been introduced primarily in
7895 context of the work to move cgroup handling to a
7896 single-writer scheme, where only PID 1
7897 creates/removes/manages cgroups.
7899 * There's a new concept of "transient" units. In contrast to
7900 normal units these units are created via an API at runtime,
7901 not from configuration from disk. More specifically this
7902 means it is now possible to run arbitrary programs as
7903 independent services, with all execution parameters passed
7904 in via bus APIs rather than read from disk. Transient units
7905 make systemd substantially more dynamic then it ever was,
7906 and useful as a general batch manager.
7908 * logind has been updated to make use of scope and slice units
7909 for managing user sessions. As a user logs in he will get
7910 his own private slice unit, to which all sessions are added
7911 as scope units. We also added support for automatically
7912 adding an instance of user@.service for the user into the
7913 slice. Effectively logind will no longer create cgroup
7914 hierarchies on its own now, it will defer entirely to PID 1
7915 for this by means of scope, service and slice units. Since
7916 user sessions this way become entities managed by PID 1
7917 the output of "systemctl" is now a lot more comprehensive.
7919 * A new mini-daemon "systemd-machined" has been added which
7920 may be used by virtualization managers to register local
7921 VMs/containers. nspawn has been updated accordingly, and
7922 libvirt will be updated shortly. machined will collect a bit
7923 of meta information about the VMs/containers, and assign
7924 them their own scope unit (see above). The collected
7925 meta-data is then made available via the "machinectl" tool,
7926 and exposed in "ps" and similar tools. machined/machinectl
7927 is compile-time optional.
7929 * As discussed earlier, the low-level cgroup configuration
7930 options ControlGroup=, ControlGroupModify=,
7931 ControlGroupPersistent=, ControlGroupAttribute= have been
7932 removed. Please use high-level attribute settings instead as
7933 well as slice units.
7935 * A new bus call SetUnitProperties() has been added to alter
7936 various runtime parameters of a unit. This is primarily
7937 useful to alter cgroup parameters dynamically in a nice way,
7938 but will be extended later on to make more properties
7939 modifiable at runtime. systemctl gained a new set-properties
7940 command that wraps this call.
7942 * A new tool "systemd-run" has been added which can be used to
7943 run arbitrary command lines as transient services or scopes,
7944 while configuring a number of settings via the command
7945 line. This tool is currently very basic, however already
7946 very useful. We plan to extend this tool to even allow
7947 queuing of execution jobs with time triggers from the
7948 command line, similar in fashion to "at".
7950 * nspawn will now inform the user explicitly that kernels with
7951 audit enabled break containers, and suggest the user to turn
7954 * Support for detecting the IMA and AppArmor security
7955 frameworks with ConditionSecurity= has been added.
7957 * journalctl gained a new "-k" switch for showing only kernel
7958 messages, mimicking dmesg output; in addition to "--user"
7959 and "--system" switches for showing only user's own logs
7962 * systemd-delta can now show information about drop-in
7963 snippets extending unit files.
7965 * libsystemd-bus has been substantially updated but is still
7966 not available as public API.
7968 * systemd will now look for the "debug" argument on the kernel
7969 command line and enable debug logging, similar to what
7970 "systemd.log_level=debug" already did before.
7972 * "systemctl set-default", "systemctl get-default" has been
7973 added to configure the default.target symlink, which
7974 controls what to boot into by default.
7976 * "systemctl set-log-level" has been added as a convenient
7977 way to raise and lower systemd logging threshold.
7979 * "systemd-analyze plot" will now show the time the various
7980 generators needed for execution, as well as information
7981 about the unit file loading.
7983 * libsystemd-journal gained a new sd_journal_open_files() call
7984 for opening specific journal files. journactl also gained a
7985 new switch to expose this new functionality. Previously we
7986 only supported opening all files from a directory, or all
7987 files from the system, as opening individual files only is
7988 racy due to journal file rotation.
7990 * systemd gained the new DefaultEnvironment= setting in
7991 /etc/systemd/system.conf to set environment variables for
7994 * If a privileged process logs a journal message with the
7995 OBJECT_PID= field set, then journald will automatically
7996 augment this with additional OBJECT_UID=, OBJECT_GID=,
7997 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
7998 system services want to log events about specific client
7999 processes. journactl/systemctl has been updated to make use
8000 of this information if all log messages regarding a specific
8003 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
8004 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
8005 Reisner, David Coppa, David King, David Strauss, Eelco
8006 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
8007 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
8008 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
8009 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
8010 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
8011 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
8012 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
8013 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
8014 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
8015 Łukasz Stelmach, 장동준
8019 * The Python bindings gained some minimal support for the APIs
8020 exposed by libsystemd-logind.
8022 * ConditionSecurity= gained support for detecting SMACK. Since
8023 this condition already supports SELinux and AppArmor we only
8024 miss IMA for this. Patches welcome!
8026 Contributions from: Karol Lewandowski, Lennart Poettering,
8027 Zbigniew Jędrzejewski-Szmek
8031 * systemd-nspawn will now create /etc/resolv.conf if
8032 necessary, before bind-mounting the host's file onto it.
8034 * systemd-nspawn will now store meta information about a
8035 container on the container's cgroup as extended attribute
8036 fields, including the root directory.
8038 * The cgroup hierarchy has been reworked in many ways. All
8039 objects any of the components systemd creates in the cgroup
8040 tree are now suffixed. More specifically, user sessions are
8041 now placed in cgroups suffixed with ".session", users in
8042 cgroups suffixed with ".user", and nspawn containers in
8043 cgroups suffixed with ".nspawn". Furthermore, all cgroup
8044 names are now escaped in a simple scheme to avoid collision
8045 of userspace object names with kernel filenames. This work
8046 is preparation for making these objects relocatable in the
8047 cgroup tree, in order to allow easy resource partitioning of
8048 these objects without causing naming conflicts.
8050 * systemctl list-dependencies gained the new switches
8051 --plain, --reverse, --after and --before.
8053 * systemd-inhibit now shows the process name of processes that
8054 have taken an inhibitor lock.
8056 * nss-myhostname will now also resolve "localhost"
8057 implicitly. This makes /etc/hosts an optional file and
8058 nicely handles that on IPv6 ::1 maps to both "localhost" and
8061 * libsystemd-logind.so gained a new call
8062 sd_get_machine_names() to enumerate running containers and
8063 VMs (currently only supported by very new libvirt and
8064 nspawn). sd_login_monitor can now be used to watch
8065 VMs/containers coming and going.
8067 * .include is not allowed recursively anymore, and only in
8068 unit files. Usually it is better to use drop-in snippets in
8069 .d/*.conf anyway, as introduced with systemd 198.
8071 * systemd-analyze gained a new "critical-chain" command that
8072 determines the slowest chain of units run during system
8073 boot-up. It is very useful for tracking down where
8074 optimizing boot time is the most beneficial.
8076 * systemd will no longer allow manipulating service paths in
8077 the name=systemd:/system cgroup tree using ControlGroup= in
8078 units. (But is still fine with it in all other dirs.)
8080 * There's a new systemd-nspawn@.service service file that may
8081 be used to easily run nspawn containers as system
8082 services. With the container's root directory in
8083 /var/lib/container/foobar it is now sufficient to run
8084 "systemctl start systemd-nspawn@foobar.service" to boot it.
8086 * systemd-cgls gained a new parameter "--machine" to list only
8087 the processes within a certain container.
8089 * ConditionSecurity= now can check for "apparmor". We still
8090 are lacking checks for SMACK and IMA for this condition
8091 check though. Patches welcome!
8093 * A new configuration file /etc/systemd/sleep.conf has been
8094 added that may be used to configure which kernel operation
8095 systemd is supposed to execute when "suspend", "hibernate"
8096 or "hybrid-sleep" is requested. This makes the new kernel
8097 "freeze" state accessible to the user.
8099 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
8100 the passed argument if applicable.
8102 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
8103 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
8104 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
8105 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
8106 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
8107 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
8108 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
8113 * The output of 'systemctl list-jobs' got some polishing. The
8114 '--type=' argument may now be passed more than once. A new
8115 command 'systemctl list-sockets' has been added which shows
8116 a list of kernel sockets systemd is listening on with the
8117 socket units they belong to, plus the units these socket
8120 * The experimental libsystemd-bus library got substantial
8121 updates to work in conjunction with the (also experimental)
8122 kdbus kernel project. It works well enough to exchange
8123 messages with some sophistication. Note that kdbus is not
8124 ready yet, and the library is mostly an elaborate test case
8125 for now, and not installable.
8127 * systemd gained a new unit 'systemd-static-nodes.service'
8128 that generates static device nodes earlier during boot, and
8129 can run in conjunction with udev.
8131 * libsystemd-login gained a new call sd_pid_get_user_unit()
8132 to retrieve the user systemd unit a process is running
8133 in. This is useful for systems where systemd is used as
8136 * systemd-nspawn now places all containers in the new /machine
8137 top-level cgroup directory in the name=systemd
8138 hierarchy. libvirt will soon do the same, so that we get a
8139 uniform separation of /system, /user and /machine for system
8140 services, user processes and containers/virtual
8141 machines. This new cgroup hierarchy is also useful to stick
8142 stable names to specific container instances, which can be
8143 recognized later this way (this name may be controlled
8144 via systemd-nspawn's new -M switch). libsystemd-login also
8145 gained a new call sd_pid_get_machine_name() to retrieve the
8146 name of the container/VM a specific process belongs to.
8148 * bootchart can now store its data in the journal.
8150 * libsystemd-journal gained a new call
8151 sd_journal_add_conjunction() for AND expressions to the
8152 matching logic. This can be used to express more complex
8153 logical expressions.
8155 * journactl can now take multiple --unit= and --user-unit=
8158 * The cryptsetup logic now understands the "luks.key=" kernel
8159 command line switch for specifying a file to read the
8160 decryption key from. Also, if a configured key file is not
8161 found the tool will now automatically fall back to prompting
8164 * Python systemd.journal module was updated to wrap recently
8165 added functions from libsystemd-journal. The interface was
8166 changed to bring the low level interface in s.j._Reader
8167 closer to the C API, and the high level interface in
8168 s.j.Reader was updated to wrap and convert all data about
8171 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
8172 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
8173 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
8174 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
8175 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
8176 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8180 * journalctl --update-catalog now understands a new --root=
8181 option to operate on catalogs found in a different root
8184 * During shutdown after systemd has terminated all running
8185 services a final killing loop kills all remaining left-over
8186 processes. We will now print the name of these processes
8187 when we send SIGKILL to them, since this usually indicates a
8190 * If /etc/crypttab refers to password files stored on
8191 configured mount points automatic dependencies will now be
8192 generated to ensure the specific mount is established first
8193 before the key file is attempted to be read.
8195 * 'systemctl status' will now show information about the
8196 network sockets a socket unit is listening on.
8198 * 'systemctl status' will also shown information about any
8199 drop-in configuration file for units. (Drop-In configuration
8200 files in this context are files such as
8201 /etc/systemd/systemd/foobar.service.d/*.conf)
8203 * systemd-cgtop now optionally shows summed up CPU times of
8204 cgroups. Press '%' while running cgtop to switch between
8205 percentage and absolute mode. This is useful to determine
8206 which cgroups use up the most CPU time over the entire
8207 runtime of the system. systemd-cgtop has also been updated
8208 to be 'pipeable' for processing with further shell tools.
8210 * 'hostnamectl set-hostname' will now allow setting of FQDN
8213 * The formatting and parsing of time span values has been
8214 changed. The parser now understands fractional expressions
8215 such as "5.5h". The formatter will now output fractional
8216 expressions for all time spans under 1min, i.e. "5.123456s"
8217 rather than "5s 123ms 456us". For time spans under 1s
8218 millisecond values are shown, for those under 1ms
8219 microsecond values are shown. This should greatly improve
8220 all time-related output of systemd.
8222 * libsystemd-login and libsystemd-journal gained new
8223 functions for querying the poll() events mask and poll()
8224 timeout value for integration into arbitrary event
8227 * localectl gained the ability to list available X11 keymaps
8228 (models, layouts, variants, options).
8230 * 'systemd-analyze dot' gained the ability to filter for
8231 specific units via shell-style globs, to create smaller,
8232 more useful graphs. I.e. it is now possible to create simple
8233 graphs of all the dependencies between only target units, or
8234 of all units that Avahi has dependencies with.
8236 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
8237 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
8238 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
8239 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
8240 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
8241 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
8242 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
8246 * The boot-time readahead implementation for rotating media
8247 will now read the read-ahead data in multiple passes which
8248 consist of all read requests made in equidistant time
8249 intervals. This means instead of strictly reading read-ahead
8250 data in its physical order on disk we now try to find a
8251 middle ground between physical and access time order.
8253 * /etc/os-release files gained a new BUILD_ID= field for usage
8254 on operating systems that provide continuous builds of OS
8257 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
8258 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
8259 William Douglas, Zbigniew Jędrzejewski-Szmek
8263 * systemd-python gained an API exposing libsystemd-daemon.
8265 * The SMACK setup logic gained support for uploading CIPSO
8268 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
8269 ReadOnlyDirectories= and InaccessibleDirectories= has
8270 changed. The private /tmp and /var/tmp directories are now
8271 shared by all processes of a service (which means
8272 ExecStartPre= may now leave data in /tmp that ExecStart= of
8273 the same service can still access). When a service is
8274 stopped its temporary directories are immediately deleted
8275 (normal clean-up with tmpfiles is still done in addition to
8278 * By default, systemd will now set a couple of sysctl
8279 variables in the kernel: the safe sysrq options are turned
8280 on, IP route verification is turned on, and source routing
8281 disabled. The recently added hardlink and softlink
8282 protection of the kernel is turned on. These settings should
8283 be reasonably safe, and good defaults for all new systems.
8285 * The predictable network naming logic may now be turned off
8286 with a new kernel command line switch: net.ifnames=0.
8288 * A new libsystemd-bus module has been added that implements a
8289 pretty complete D-Bus client library. For details see:
8291 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
8293 * journald will now explicitly flush the journal files to disk
8294 at the latest 5min after each write. The file will then also
8295 be marked offline until the next write. This should increase
8296 reliability in case of a crash. The synchronization delay
8297 can be configured via SyncIntervalSec= in journald.conf.
8299 * There's a new remote-fs-setup.target unit that can be used
8300 to pull in specific services when at least one remote file
8301 system is to be mounted.
8303 * There are new targets timers.target and paths.target as
8304 canonical targets to pull user timer and path units in
8305 from. This complements sockets.target with a similar
8306 purpose for socket units.
8308 * libudev gained a new call udev_device_set_attribute_value()
8309 to set sysfs attributes of a device.
8311 * The udev daemon now sets the default number of worker
8312 processes executed in parallel based on the number of available
8313 CPUs instead of the amount of available RAM. This is supposed
8314 to provide a more reliable default and limit a too aggressive
8315 parallelism for setups with 1000s of devices connected.
8317 Contributions from: Auke Kok, Colin Walters, Cristian
8318 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
8319 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
8320 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
8321 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
8322 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
8323 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
8324 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
8325 Zbigniew Jędrzejewski-Szmek
8329 * Configuration of unit files may now be extended via drop-in
8330 files without having to edit/override the unit files
8331 themselves. More specifically, if the administrator wants to
8332 change one value for a service file foobar.service he can
8333 now do so by dropping in a configuration snippet into
8334 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
8335 will load all these snippets and apply them on top of the
8336 main unit configuration file, possibly extending or
8337 overriding its settings. Using these drop-in snippets is
8338 generally nicer than the two earlier options for changing
8339 unit files locally: copying the files from
8340 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
8341 them there; or creating a new file in /etc/systemd/system/
8342 that incorporates the original one via ".include". Drop-in
8343 snippets into these .d/ directories can be placed in any
8344 directory systemd looks for units in, and the usual
8345 overriding semantics between /usr/lib, /etc and /run apply
8348 * Most unit file settings which take lists of items can now be
8349 reset by assigning the empty string to them. For example,
8350 normally, settings such as Environment=FOO=BAR append a new
8351 environment variable assignment to the environment block,
8352 each time they are used. By assigning Environment= the empty
8353 string the environment block can be reset to empty. This is
8354 particularly useful with the .d/*.conf drop-in snippets
8355 mentioned above, since this adds the ability to reset list
8356 settings from vendor unit files via these drop-ins.
8358 * systemctl gained a new "list-dependencies" command for
8359 listing the dependencies of a unit recursively.
8361 * Inhibitors are now honored and listed by "systemctl
8362 suspend", "systemctl poweroff" (and similar) too, not only
8363 GNOME. These commands will also list active sessions by
8366 * Resource limits (as exposed by the various control group
8367 controllers) can now be controlled dynamically at runtime
8368 for all units. More specifically, you can now use a command
8369 like "systemctl set-cgroup-attr foobar.service cpu.shares
8370 2000" to alter the CPU shares a specific service gets. These
8371 settings are stored persistently on disk, and thus allow the
8372 administrator to easily adjust the resource usage of
8373 services with a few simple commands. This dynamic resource
8374 management logic is also available to other programs via the
8375 bus. Almost any kernel cgroup attribute and controller is
8378 * systemd-vconsole-setup will now copy all font settings to
8379 all allocated VTs, where it previously applied them only to
8382 * libsystemd-login gained the new sd_session_get_tty() API
8385 * This release drops support for a few legacy or
8386 distribution-specific LSB facility names when parsing init
8387 scripts: $x-display-manager, $mail-transfer-agent,
8388 $mail-transport-agent, $mail-transfer-agent, $smtp,
8389 $null. Also, the mail-transfer-agent.target unit backing
8390 this has been removed. Distributions which want to retain
8391 compatibility with this should carry the burden for
8392 supporting this themselves and patch support for these back
8393 in, if they really need to. Also, the facilities $syslog and
8394 $local_fs are now ignored, since systemd does not support
8395 early-boot LSB init scripts anymore, and these facilities
8396 are implied anyway for normal services. syslog.target has
8399 * There are new bus calls on PID1's Manager object for
8400 cancelling jobs, and removing snapshot units. Previously,
8401 both calls were only available on the Job and Snapshot
8404 * systemd-journal-gatewayd gained SSL support.
8406 * The various "environment" files, such as /etc/locale.conf
8407 now support continuation lines with a backslash ("\") as
8408 last character in the line, similarly in style (but different)
8409 to how this is supported in shells.
8411 * For normal user processes the _SYSTEMD_USER_UNIT= field is
8412 now implicitly appended to every log entry logged. systemctl
8413 has been updated to filter by this field when operating on a
8414 user systemd instance.
8416 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
8417 CAP_AUDIT_CONTROL capabilities to the capabilities set for
8418 the container. This makes it easier to boot unmodified
8419 Fedora systems in a container, which however still requires
8420 audit=0 to be passed on the kernel command line. Auditing in
8421 kernel and userspace is unfortunately still too broken in
8422 context of containers, hence we recommend compiling it out
8423 of the kernel or using audit=0. Hopefully this will be fixed
8424 one day for good in the kernel.
8426 * nspawn gained the new --bind= and --bind-ro= parameters to
8427 bind mount specific directories from the host into the
8430 * nspawn will now mount its own devpts file system instance
8431 into the container, in order not to leak pty devices from
8432 the host into the container.
8434 * systemd will now read the firmware boot time performance
8435 information from the EFI variables, if the used boot loader
8436 supports this, and takes it into account for boot performance
8437 analysis via "systemd-analyze". This is currently supported
8438 only in conjunction with Gummiboot, but could be supported
8439 by other boot loaders too. For details see:
8441 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
8443 * A new generator has been added that automatically mounts the
8444 EFI System Partition (ESP) to /boot, if that directory
8445 exists, is empty, and no other file system has been
8446 configured to be mounted there.
8448 * logind will now send out PrepareForSleep(false) out
8449 unconditionally, after coming back from suspend. This may be
8450 used by applications as asynchronous notification for
8451 system resume events.
8453 * "systemctl unlock-sessions" has been added, that allows
8454 unlocking the screens of all user sessions at once, similar
8455 to how "systemctl lock-sessions" already locked all users
8456 sessions. This is backed by a new D-Bus call UnlockSessions().
8458 * "loginctl seat-status" will now show the master device of a
8459 seat. (i.e. the device of a seat that needs to be around for
8460 the seat to be considered available, usually the graphics
8463 * tmpfiles gained a new "X" line type, that allows
8464 configuration of files and directories (with wildcards) that
8465 shall be excluded from automatic cleanup ("aging").
8467 * udev default rules set the device node permissions now only
8468 at "add" events, and do not change them any longer with a
8469 later "change" event.
8471 * The log messages for lid events and power/sleep keypresses
8472 now carry a message ID.
8474 * We now have a substantially larger unit test suite, but this
8475 continues to be work in progress.
8477 * udevadm hwdb gained a new --root= parameter to change the
8478 root directory to operate relative to.
8480 * logind will now issue a background sync() request to the kernel
8481 early at shutdown, so that dirty buffers are flushed to disk early
8482 instead of at the last moment, in order to optimize shutdown
8485 * A new bootctl tool has been added that is an interface for
8486 certain boot loader operations. This is currently a preview
8487 and is likely to be extended into a small mechanism daemon
8488 like timedated, localed, hostnamed, and can be used by
8489 graphical UIs to enumerate available boot options, and
8490 request boot into firmware operations.
8492 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
8493 the rest of the package. It also has been updated to work
8494 correctly in initrds.
8496 * polkit previously has been runtime optional, and is now also
8497 compile time optional via a configure switch.
8499 * systemd-analyze has been reimplemented in C. Also "systemctl
8500 dot" has moved into systemd-analyze.
8502 * "systemctl status" with no further parameters will now print
8503 the status of all active or failed units.
8505 * Operations such as "systemctl start" can now be executed
8506 with a new mode "--irreversible" which may be used to queue
8507 operations that cannot accidentally be reversed by a later
8508 job queuing. This is by default used to make shutdown
8509 requests more robust.
8511 * The Python API of systemd now gained a new module for
8512 reading journal files.
8514 * A new tool kernel-install has been added that can install
8515 kernel images according to the Boot Loader Specification:
8517 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
8519 * Boot time console output has been improved to provide
8520 animated boot time output for hanging jobs.
8522 * A new tool systemd-activate has been added which can be used
8523 to test socket activation with, directly from the command
8524 line. This should make it much easier to test and debug
8525 socket activation in daemons.
8527 * journalctl gained a new "--reverse" (or -r) option to show
8528 journal output in reverse order (i.e. newest line first).
8530 * journalctl gained a new "--pager-end" (or -e) option to jump
8531 to immediately jump to the end of the journal in the
8532 pager. This is only supported in conjunction with "less".
8534 * journalctl gained a new "--user-unit=" option, that works
8535 similarly to "--unit=" but filters for user units rather than
8538 * A number of unit files to ease adoption of systemd in
8539 initrds has been added. This moves some minimal logic from
8540 the various initrd implementations into systemd proper.
8542 * The journal files are now owned by a new group
8543 "systemd-journal", which exists specifically to allow access
8544 to the journal, and nothing else. Previously, we used the
8545 "adm" group for that, which however possibly covers more
8546 than just journal/log file access. This new group is now
8547 already used by systemd-journal-gatewayd to ensure this
8548 daemon gets access to the journal files and as little else
8549 as possible. Note that "make install" will also set FS ACLs
8550 up for /var/log/journal to give "adm" and "wheel" read
8551 access to it, in addition to "systemd-journal" which owns
8552 the journal files. We recommend that packaging scripts also
8553 add read access to "adm" + "wheel" to /var/log/journal, and
8554 all existing/future journal files. To normal users and
8555 administrators little changes, however packagers need to
8556 ensure to create the "systemd-journal" system group at
8557 package installation time.
8559 * The systemd-journal-gatewayd now runs as unprivileged user
8560 systemd-journal-gateway:systemd-journal-gateway. Packaging
8561 scripts need to create these system user/group at
8564 * timedated now exposes a new boolean property CanNTP that
8565 indicates whether a local NTP service is available or not.
8567 * systemd-detect-virt will now also detect xen PVs
8569 * The pstore file system is now mounted by default, if it is
8572 * In addition to the SELinux and IMA policies we will now also
8573 load SMACK policies at early boot.
8575 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
8576 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
8577 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
8578 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
8579 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
8580 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
8581 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
8582 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
8583 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
8584 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
8585 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
8586 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
8587 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
8588 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
8592 * Timer units now support calendar time events in addition to
8593 monotonic time events. That means you can now trigger a unit
8594 based on a calendar time specification such as "Thu,Fri
8595 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
8596 or fifth day of any month of the year 2013, given that it is
8597 a thursday or friday. This brings timer event support
8598 considerably closer to cron's capabilities. For details on
8599 the supported calendar time specification language see
8602 * udev now supports a number of different naming policies for
8603 network interfaces for predictable names, and a combination
8604 of these policies is now the default. Please see this wiki
8605 document for details:
8607 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8609 * Auke Kok's bootchart implementation has been added to the
8610 systemd tree. It is an optional component that can graph the
8611 boot in quite some detail. It is one of the best bootchart
8612 implementations around and minimal in its code and
8615 * nss-myhostname has been integrated into the systemd source
8616 tree. nss-myhostname guarantees that the local hostname
8617 always stays resolvable via NSS. It has been a weak
8618 requirement of systemd-hostnamed since a long time, and
8619 since its code is actually trivial we decided to just
8620 include it in systemd's source tree. It can be turned off
8621 with a configure switch.
8623 * The read-ahead logic is now capable of properly detecting
8624 whether a btrfs file system is on SSD or rotating media, in
8625 order to optimize the read-ahead scheme. Previously, it was
8626 only capable of detecting this on traditional file systems
8629 * In udev, additional device properties are now read from the
8630 IAB in addition to the OUI database. Also, Bluetooth company
8631 identities are attached to the devices as well.
8633 * In service files %U may be used as specifier that is
8634 replaced by the configured user name of the service.
8636 * nspawn may now be invoked without a controlling TTY. This
8637 makes it suitable for invocation as its own service. This
8638 may be used to set up a simple containerized server system
8639 using only core OS tools.
8641 * systemd and nspawn can now accept socket file descriptors
8642 when they are started for socket activation. This enables
8643 implementation of socket activated nspawn
8644 containers. i.e. think about autospawning an entire OS image
8645 when the first SSH or HTTP connection is received. We expect
8646 that similar functionality will also be added to libvirt-lxc
8649 * journalctl will now suppress ANSI color codes when
8650 presenting log data.
8652 * systemctl will no longer show control group information for
8653 a unit if the control group is empty anyway.
8655 * logind can now automatically suspend/hibernate/shutdown the
8658 * /etc/machine-info and hostnamed now also expose the chassis
8659 type of the system. This can be used to determine whether
8660 the local system is a laptop, desktop, handset or
8661 tablet. This information may either be configured by the
8662 user/vendor or is automatically determined from ACPI and DMI
8663 information if possible.
8665 * A number of polkit actions are now bound together with "imply"
8666 rules. This should simplify creating UIs because many actions
8667 will now authenticate similar ones as well.
8669 * Unit files learnt a new condition ConditionACPower= which
8670 may be used to conditionalize a unit depending on whether an
8671 AC power source is connected or not, of whether the system
8672 is running on battery power.
8674 * systemctl gained a new "is-failed" verb that may be used in
8675 shell scripts and suchlike to check whether a specific unit
8676 is in the "failed" state.
8678 * The EnvironmentFile= setting in unit files now supports file
8679 globbing, and can hence be used to easily read a number of
8680 environment files at once.
8682 * systemd will no longer detect and recognize specific
8683 distributions. All distribution-specific #ifdeffery has been
8684 removed, systemd is now fully generic and
8685 distribution-agnostic. Effectively, not too much is lost as
8686 a lot of the code is still accessible via explicit configure
8687 switches. However, support for some distribution specific
8688 legacy configuration file formats has been dropped. We
8689 recommend distributions to simply adopt the configuration
8690 files everybody else uses now and convert the old
8691 configuration from packaging scripts. Most distributions
8692 already did that. If that's not possible or desirable,
8693 distributions are welcome to forward port the specific
8694 pieces of code locally from the git history.
8696 * When logging a message about a unit systemd will now always
8697 log the unit name in the message meta data.
8699 * localectl will now also discover system locale data that is
8700 not stored in locale archives, but directly unpacked.
8702 * logind will no longer unconditionally use framebuffer
8703 devices as seat masters, i.e. as devices that are required
8704 to be existing before a seat is considered preset. Instead,
8705 it will now look for all devices that are tagged as
8706 "seat-master" in udev. By default, framebuffer devices will
8707 be marked as such, but depending on local systems, other
8708 devices might be marked as well. This may be used to
8709 integrate graphics cards using closed source drivers (such
8710 as NVidia ones) more nicely into logind. Note however, that
8711 we recommend using the open source NVidia drivers instead,
8712 and no udev rules for the closed-source drivers will be
8713 shipped from us upstream.
8715 Contributions from: Adam Williamson, Alessandro Crismani, Auke
8716 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
8717 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
8718 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
8719 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
8720 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
8721 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
8722 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
8723 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
8724 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
8725 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
8726 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
8731 * udev gained support for loading additional device properties
8732 from an indexed database that is keyed by vendor/product IDs
8733 and similar device identifiers. For the beginning this
8734 "hwdb" is populated with data from the well-known PCI and
8735 USB database, but also includes PNP, ACPI and OID data. In
8736 the longer run this indexed database shall grow into
8737 becoming the one central database for non-essential
8738 userspace device metadata. Previously, data from the PCI/USB
8739 database was only attached to select devices, since the
8740 lookup was a relatively expensive operation due to O(n) time
8741 complexity (with n being the number of entries in the
8742 database). Since this is now O(1), we decided to add in this
8743 data for all devices where this is available, by
8744 default. Note that the indexed database needs to be rebuilt
8745 when new data files are installed. To achieve this you need
8746 to update your packaging scripts to invoke "udevadm hwdb
8747 --update" after installation of hwdb data files. For
8748 RPM-based distributions we introduced the new
8749 %udev_hwdb_update macro for this purpose.
8751 * The Journal gained support for the "Message Catalog", an
8752 indexed database to link up additional information with
8753 journal entries. For further details please check:
8755 https://www.freedesktop.org/wiki/Software/systemd/catalog
8757 The indexed message catalog database also needs to be
8758 rebuilt after installation of message catalog files. Use
8759 "journalctl --update-catalog" for this. For RPM-based
8760 distributions we introduced the %journal_catalog_update
8761 macro for this purpose.
8763 * The Python Journal bindings gained support for the standard
8764 Python logging framework.
8766 * The Journal API gained new functions for checking whether
8767 the underlying file system of a journal file is capable of
8768 properly reporting file change notifications, or whether
8769 applications that want to reflect journal changes "live"
8770 need to recheck journal files continuously in appropriate
8773 * It is now possible to set the "age" field for tmpfiles
8774 entries to 0, indicating that files matching this entry
8775 shall always be removed when the directories are cleaned up.
8777 * coredumpctl gained a new "gdb" verb which invokes gdb
8778 right-away on the selected coredump.
8780 * There's now support for "hybrid sleep" on kernels that
8781 support this, in addition to "suspend" and "hibernate". Use
8782 "systemctl hybrid-sleep" to make use of this.
8784 * logind's HandleSuspendKey= setting (and related settings)
8785 now gained support for a new "lock" setting to simply
8786 request the screen lock on all local sessions, instead of
8787 actually executing a suspend or hibernation.
8789 * systemd will now mount the EFI variables file system by
8792 * Socket units now gained support for configuration of the
8793 SMACK security label.
8795 * timedatectl will now output the time of the last and next
8796 daylight saving change.
8798 * We dropped support for various legacy and distro-specific
8799 concepts, such as insserv, early-boot SysV services
8800 (i.e. those for non-standard runlevels such as 'b' or 'S')
8801 or ArchLinux /etc/rc.conf support. We recommend the
8802 distributions who still need support this to either continue
8803 to maintain the necessary patches downstream, or find a
8804 different solution. (Talk to us if you have questions!)
8806 * Various systemd components will now bypass polkit checks for
8807 root and otherwise handle properly if polkit is not found to
8808 be around. This should fix most issues for polkit-less
8809 systems. Quite frankly this should have been this way since
8810 day one. It is absolutely our intention to make systemd work
8811 fine on polkit-less systems, and we consider it a bug if
8812 something does not work as it should if polkit is not around.
8814 * For embedded systems it is now possible to build udev and
8815 systemd without blkid and/or kmod support.
8817 * "systemctl switch-root" is now capable of switching root
8818 more than once. I.e. in addition to transitions from the
8819 initrd to the host OS it is now possible to transition to
8820 further OS images from the host. This is useful to implement
8821 offline updating tools.
8823 * Various other additions have been made to the RPM macros
8824 shipped with systemd. Use %udev_rules_update() after
8825 installing new udev rules files. %_udevhwdbdir,
8826 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
8827 %_sysctldir are now available which resolve to the right
8828 directories for packages to place various data files in.
8830 * journalctl gained the new --full switch (in addition to
8831 --all, to disable ellipsation for long messages.
8833 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
8834 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
8835 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
8836 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
8837 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
8838 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
8839 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
8840 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
8841 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
8845 * journalctl gained new --since= and --until= switches to
8846 filter by time. It also now supports nice filtering for
8847 units via --unit=/-u.
8849 * Type=oneshot services may use ExecReload= and do the
8852 * The journal daemon now supports time-based rotation and
8853 vacuuming, in addition to the usual disk-space based
8856 * The journal will now index the available field values for
8857 each field name. This enables clients to show pretty drop
8858 downs of available match values when filtering. The bash
8859 completion of journalctl has been updated
8860 accordingly. journalctl gained a new switch -F to list all
8861 values a certain field takes in the journal database.
8863 * More service events are now written as structured messages
8864 to the journal, and made recognizable via message IDs.
8866 * The timedated, localed and hostnamed mini-services which
8867 previously only provided support for changing time, locale
8868 and hostname settings from graphical DEs such as GNOME now
8869 also have a minimal (but very useful) text-based client
8870 utility each. This is probably the nicest way to changing
8871 these settings from the command line now, especially since
8872 it lists available options and is fully integrated with bash
8875 * There's now a new tool "systemd-coredumpctl" to list and
8876 extract coredumps from the journal.
8878 * We now install a README each in /var/log/ and
8879 /etc/rc.d/init.d explaining where the system logs and init
8880 scripts went. This hopefully should help folks who go to
8881 that dirs and look into the otherwise now empty void and
8882 scratch their heads.
8884 * When user-services are invoked (by systemd --user) the
8885 $MANAGERPID env var is set to the PID of systemd.
8887 * SIGRTMIN+24 when sent to a --user instance will now result
8888 in immediate termination of systemd.
8890 * gatewayd received numerous feature additions such as a
8891 "follow" mode, for live syncing and filtering.
8893 * browse.html now allows filtering and showing detailed
8894 information on specific entries. Keyboard navigation and
8895 mouse screen support has been added.
8897 * gatewayd/journalctl now supports HTML5/JSON
8898 Server-Sent-Events as output.
8900 * The SysV init script compatibility logic will now
8901 heuristically determine whether a script supports the
8902 "reload" verb, and only then make this available as
8905 * "systemctl status --follow" has been removed, use "journalctl
8908 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
8909 have been removed since they are hardly useful to be
8912 * And I'd like to take the opportunity to specifically mention
8913 Zbigniew for his great contributions. Zbigniew, you rock!
8915 Contributions from: Andrew Eikum, Christian Hesse, Colin
8916 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
8917 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
8918 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
8919 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
8920 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
8921 Jędrzejewski-Szmek, Сковорода Никита Андреевич
8925 * If /etc/vconsole.conf is non-existent or empty we will no
8926 longer load any console font or key map at boot by
8927 default. Instead the kernel defaults will be left
8928 intact. This is definitely the right thing to do, as no
8929 configuration should mean no configuration, and hard-coding
8930 font names that are different on all archs is probably a bad
8931 idea. Also, the kernel default key map and font should be
8932 good enough for most cases anyway, and mostly identical to
8933 the userspace fonts/key maps we previously overloaded them
8934 with. If distributions want to continue to default to a
8935 non-kernel font or key map they should ship a default
8936 /etc/vconsole.conf with the appropriate contents.
8938 Contributions from: Colin Walters, Daniel J Walsh, Dave
8939 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
8940 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8944 * journalctl gained a new --cursor= switch to show entries
8945 starting from the specified location in the journal.
8947 * We now enforce a size limit on journal entry fields exported
8948 with "-o json" in journalctl. Fields larger than 4K will be
8949 assigned null. This can be turned off with --all.
8951 * An (optional) journal gateway daemon is now available as
8952 "systemd-journal-gatewayd.service". This service provides
8953 access to the journal via HTTP and JSON. This functionality
8954 will be used to implement live log synchronization in both
8955 pull and push modes, but has various other users too, such
8956 as easy log access for debugging of embedded devices. Right
8957 now it is already useful to retrieve the journal via HTTP:
8959 # systemctl start systemd-journal-gatewayd.service
8960 # wget http://localhost:19531/entries
8962 This will download the journal contents in a
8963 /var/log/messages compatible format. The same as JSON:
8965 # curl -H"Accept: application/json" http://localhost:19531/entries
8967 This service is also accessible via a web browser where a
8968 single static HTML5 app is served that uses the JSON logic
8969 to enable the user to do some basic browsing of the
8970 journal. This will be extended later on. Here's an example
8971 screenshot of this app in its current state:
8973 http://0pointer.de/public/journal-gatewayd
8975 Contributions from: Kay Sievers, Lennart Poettering, Robert
8976 Milasan, Tom Gundersen
8980 * The bash completion logic is now available for journalctl
8983 * We do not mount the "cpuset" controller anymore together with
8984 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
8985 started if no parameters are assigned to it. "cpuset" hence
8986 broke code that assumed it could create "cpu" groups and
8989 * journalctl -f will now subscribe to terminal size changes,
8990 and line break accordingly.
8992 Contributions from: Dave Reisner, Kay Sievers, Lennart
8993 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
8997 * nspawn will now create a symlink /etc/localtime in the
8998 container environment, copying the host's timezone
8999 setting. Previously this has been done via a bind mount, but
9000 since symlinks cannot be bind mounted this has now been
9001 changed to create/update the appropriate symlink.
9003 * journalctl -n's line number argument is now optional, and
9004 will default to 10 if omitted.
9006 * journald will now log the maximum size the journal files may
9007 take up on disk. This is particularly useful if the default
9008 built-in logic of determining this parameter from the file
9009 system size is used. Use "systemctl status
9010 systemd-journald.service" to see this information.
9012 * The multi-seat X wrapper tool has been stripped down. As X
9013 is now capable of enumerating graphics devices via udev in a
9014 seat-aware way the wrapper is not strictly necessary
9015 anymore. A stripped down temporary stop-gap is still shipped
9016 until the upstream display managers have been updated to
9017 fully support the new X logic. Expect this wrapper to be
9018 removed entirely in one of the next releases.
9020 * HandleSleepKey= in logind.conf has been split up into
9021 HandleSuspendKey= and HandleHibernateKey=. The old setting
9022 is not available anymore. X11 and the kernel are
9023 distinguishing between these keys and we should too. This
9024 also means the inhibition lock for these keys has been split
9027 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
9028 Poettering, Lukas Nykryn, Václav Pavlín
9032 * Whenever a unit changes state we will now log this to the
9033 journal and show along the unit's own log output in
9036 * ConditionPathIsMountPoint= can now properly detect bind
9037 mount points too. (Previously, a bind mount of one file
9038 system to another place in the same file system could not be
9039 detected as mount, since they shared struct stat's st_dev
9042 * We will now mount the cgroup controllers cpu, cpuacct,
9043 cpuset and the controllers net_cls, net_prio together by
9046 * nspawn containers will now have a virtualized boot
9047 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
9048 over with a randomized ID at container initialization). This
9049 has the effect of making "journalctl -b" do the right thing
9052 * The JSON output journal serialization has been updated not
9053 to generate "endless" list objects anymore, but rather one
9054 JSON object per line. This is more in line how most JSON
9055 parsers expect JSON objects. The new output mode
9056 "json-pretty" has been added to provide similar output, but
9057 neatly aligned for readability by humans.
9059 * We dropped all explicit sync() invocations in the shutdown
9060 code. The kernel does this implicitly anyway in the kernel
9061 reboot() syscall. halt(8)'s -n option is now a compatibility
9064 * We now support virtualized reboot() in containers, as
9065 supported by newer kernels. We will fall back to exit() if
9066 CAP_SYS_REBOOT is not available to the container. Also,
9067 nspawn makes use of this now and will actually reboot the
9068 container if the containerized OS asks for that.
9070 * journalctl will only show local log output by default
9071 now. Use --merge (-m) to show remote log output, too.
9073 * libsystemd-journal gained the new sd_journal_get_usage()
9074 call to determine the current disk usage of all journal
9075 files. This is exposed in the new "journalctl --disk-usage"
9078 * journald gained a new configuration setting SplitMode= in
9079 journald.conf which may be used to control how user journals
9080 are split off. See journald.conf(5) for details.
9082 * A new condition type ConditionFileNotEmpty= has been added.
9084 * tmpfiles' "w" lines now support file globbing, to write
9085 multiple files at once.
9087 * We added Python bindings for the journal submission
9088 APIs. More Python APIs for a number of selected APIs will
9089 likely follow. Note that we intend to add native bindings
9090 only for the Python language, as we consider it common
9091 enough to deserve bindings shipped within systemd. There are
9092 various projects outside of systemd that provide bindings
9093 for languages such as PHP or Lua.
9095 * Many conditions will now resolve specifiers such as %i. In
9096 addition, PathChanged= and related directives of .path units
9097 now support specifiers as well.
9099 * There's now a new RPM macro definition for the system preset
9102 * journald will now warn if it ca not forward a message to the
9103 syslog daemon because its socket is full.
9105 * timedated will no longer write or process /etc/timezone,
9106 except on Debian. As we do not support late mounted /usr
9107 anymore /etc/localtime always being a symlink is now safe,
9108 and hence the information in /etc/timezone is not necessary
9111 * logind will now always reserve one VT for a text getty (VT6
9112 by default). Previously if more than 6 X sessions where
9113 started they took up all the VTs with auto-spawned gettys,
9114 so that no text gettys were available anymore.
9116 * udev will now automatically inform the btrfs kernel logic
9117 about btrfs RAID components showing up. This should make
9118 simple hotplug based btrfs RAID assembly work.
9120 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
9121 (but not for its children which will stay at the kernel
9122 default). This should allow setups with a lot more listening
9125 * systemd will now always pass the configured timezone to the
9126 kernel at boot. timedated will do the same when the timezone
9129 * logind's inhibition logic has been updated. By default,
9130 logind will now handle the lid switch, the power and sleep
9131 keys all the time, even in graphical sessions. If DEs want
9132 to handle these events on their own they should take the new
9133 handle-power-key, handle-sleep-key and handle-lid-switch
9134 inhibitors during their runtime. A simple way to achieve
9135 that is to invoke the DE wrapped in an invocation of:
9137 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
9139 * Access to unit operations is now checked via SELinux taking
9140 the unit file label and client process label into account.
9142 * systemd will now notify the administrator in the journal
9143 when he over-mounts a non-empty directory.
9145 * There are new specifiers that are resolved in unit files,
9146 for the host name (%H), the machine ID (%m) and the boot ID
9149 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
9150 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
9151 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
9152 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
9153 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
9154 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
9155 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
9159 * Support for reading structured kernel messages from
9160 /dev/kmsg has now been added and is enabled by default.
9162 * Support for reading kernel messages from /proc/kmsg has now
9163 been removed. If you want kernel messages in the journal
9164 make sure to run a recent kernel (>= 3.5) that supports
9165 reading structured messages from /dev/kmsg (see
9166 above). /proc/kmsg is now exclusive property of classic
9167 syslog daemons again.
9169 * The libudev API gained the new
9170 udev_device_new_from_device_id() call.
9172 * The logic for file system namespace (ReadOnlyDirectory=,
9173 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
9174 require pivot_root() anymore. This means fewer temporary
9175 directories are created below /tmp for this feature.
9177 * nspawn containers will now see and receive all submounts
9178 made on the host OS below the root file system of the
9181 * Forward Secure Sealing is now supported for Journal files,
9182 which provide cryptographical sealing of journal files so
9183 that attackers cannot alter log history anymore without this
9184 being detectable. Lennart will soon post a blog story about
9185 this explaining it in more detail.
9187 * There are two new service settings RestartPreventExitStatus=
9188 and SuccessExitStatus= which allow configuration of exit
9189 status (exit code or signal) which will be excepted from the
9190 restart logic, resp. consider successful.
9192 * journalctl gained the new --verify switch that can be used
9193 to check the integrity of the structure of journal files and
9194 (if Forward Secure Sealing is enabled) the contents of
9197 * nspawn containers will now be run with /dev/stdin, /dev/fd/
9198 and similar symlinks pre-created. This makes running shells
9199 as container init process a lot more fun.
9201 * The fstab support can now handle PARTUUID= and PARTLABEL=
9204 * A new ConditionHost= condition has been added to match
9205 against the hostname (with globs) and machine ID. This is
9206 useful for clusters where a single OS image is used to
9207 provision a large number of hosts which shall run slightly
9208 different sets of services.
9210 * Services which hit the restart limit will now be placed in a
9213 Contributions from: Bertram Poettering, Dave Reisner, Huang
9214 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
9215 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
9219 * When running in --user mode systemd will now become a
9220 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
9221 tree a lot more organized.
9223 * A new PartOf= unit dependency type has been introduced that
9224 may be used to group services in a natural way.
9226 * "systemctl enable" may now be used to enable instances of
9229 * journalctl now prints error log levels in red, and
9230 warning/notice log levels in bright white. It also supports
9231 filtering by log level now.
9233 * cgtop gained a new -n switch (similar to top), to configure
9234 the maximum number of iterations to run for. It also gained
9235 -b, to run in batch mode (accepting no input).
9237 * The suffix ".service" may now be omitted on most systemctl
9238 command lines involving service unit names.
9240 * There's a new bus call in logind to lock all sessions, as
9241 well as a loginctl verb for it "lock-sessions".
9243 * libsystemd-logind.so gained a new call sd_journal_perror()
9244 that works similar to libc perror() but logs to the journal
9245 and encodes structured information about the error number.
9247 * /etc/crypttab entries now understand the new keyfile-size=
9250 * shutdown(8) now can send a (configurable) wall message when
9251 a shutdown is cancelled.
9253 * The mount propagation mode for the root file system will now
9254 default to "shared", which is useful to make containers work
9255 nicely out-of-the-box so that they receive new mounts from
9256 the host. This can be undone locally by running "mount
9257 --make-rprivate /" if needed.
9259 * The prefdm.service file has been removed. Distributions
9260 should maintain this unit downstream if they intend to keep
9261 it around. However, we recommend writing normal unit files
9262 for display managers instead.
9264 * Since systemd is a crucial part of the OS we will now
9265 default to a number of compiler switches that improve
9266 security (hardening) such as read-only relocations, stack
9267 protection, and suchlike.
9269 * The TimeoutSec= setting for services is now split into
9270 TimeoutStartSec= and TimeoutStopSec= to allow configuration
9271 of individual time outs for the start and the stop phase of
9274 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
9275 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
9276 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
9277 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
9278 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
9279 Gundersen, Zbigniew Jędrzejewski-Szmek
9283 * The journal and id128 C APIs are now fully documented as man
9286 * Extra safety checks have been added when transitioning from
9287 the initial RAM disk to the main system to avoid accidental
9290 * /etc/crypttab entries now understand the new keyfile-offset=
9293 * systemctl -t can now be used to filter by unit load state.
9295 * The journal C API gained the new sd_journal_wait() call to
9296 make writing synchronous journal clients easier.
9298 * journalctl gained the new -D switch to show journals from a
9301 * journalctl now displays a special marker between log
9302 messages of two different boots.
9304 * The journal is now explicitly flushed to /var via a service
9305 systemd-journal-flush.service, rather than implicitly simply
9306 by seeing /var/log/journal to be writable.
9308 * journalctl (and the journal C APIs) can now match for much
9309 more complex expressions, with alternatives and
9312 * When transitioning from the initial RAM disk to the main
9313 system we will now kill all processes in a killing spree to
9314 ensure no processes stay around by accident.
9316 * Three new specifiers may be used in unit files: %u, %h, %s
9317 resolve to the user name, user home directory resp. user
9318 shell. This is useful for running systemd user instances.
9320 * We now automatically rotate journal files if their data
9321 object hash table gets a fill level > 75%. We also size the
9322 hash table based on the configured maximum file size. This
9323 together should lower hash collisions drastically and thus
9324 speed things up a bit.
9326 * journalctl gained the new "--header" switch to introspect
9327 header data of journal files.
9329 * A new setting SystemCallFilters= has been added to services
9330 which may be used to apply blacklists or whitelists to
9331 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
9333 * nspawn gained a new --link-journal= switch (and quicker: -j)
9334 to link the container journal with the host. This makes it
9335 very easy to centralize log viewing on the host for all
9336 guests while still keeping the journal files separated.
9338 * Many bugfixes and optimizations
9340 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
9341 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
9342 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
9347 * Several tools now understand kernel command line arguments,
9348 which are only read when run in an initial RAM disk. They
9349 usually follow closely their normal counterparts, but are
9352 * There's a new tool to analyze the readahead files that are
9353 automatically generated at boot. Use:
9355 /usr/lib/systemd/systemd-readahead analyze /.readahead
9357 * We now provide an early debug shell on tty9 if this enabled. Use:
9359 systemctl enable debug-shell.service
9361 * All plymouth related units have been moved into the Plymouth
9362 package. Please make sure to upgrade your Plymouth version
9365 * systemd-tmpfiles now supports getting passed the basename of
9366 a configuration file only, in which case it will look for it
9367 in all appropriate directories automatically.
9369 * udevadm info now takes a /dev or /sys path as argument, and
9370 does the right thing. Example:
9372 udevadm info /dev/sda
9373 udevadm info /sys/class/block/sda
9375 * systemctl now prints a warning if a unit is stopped but a
9376 unit that might trigger it continues to run. Example: a
9377 service is stopped but the socket that activates it is left
9380 * "systemctl status" will now mention if the log output was
9381 shortened due to rotation since a service has been started.
9383 * The journal API now exposes functions to determine the
9384 "cutoff" times due to rotation.
9386 * journald now understands SIGUSR1 and SIGUSR2 for triggering
9387 immediately flushing of runtime logs to /var if possible,
9388 resp. for triggering immediate rotation of the journal
9391 * It is now considered an error if a service is attempted to
9392 be stopped that is not loaded.
9394 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
9396 * systemd-analyze now supports Python 3
9398 * tmpfiles now supports cleaning up directories via aging
9399 where the first level dirs are always kept around but
9400 directories beneath it automatically aged. This is enabled
9401 by prefixing the age field with '~'.
9403 * Seat objects now expose CanGraphical, CanTTY properties
9404 which is required to deal with very fast bootups where the
9405 display manager might be running before the graphics drivers
9406 completed initialization.
9408 * Seat objects now expose a State property.
9410 * We now include RPM macros for service enabling/disabling
9411 based on the preset logic. We recommend RPM based
9412 distributions to make use of these macros if possible. This
9413 makes it simpler to reuse RPM spec files across
9416 * We now make sure that the collected systemd unit name is
9417 always valid when services log to the journal via
9420 * There's a new man page kernel-command-line(7) detailing all
9421 command line options we understand.
9423 * The fstab generator may now be disabled at boot by passing
9424 fstab=0 on the kernel command line.
9426 * A new kernel command line option modules-load= is now understood
9427 to load a specific kernel module statically, early at boot.
9429 * Unit names specified on the systemctl command line are now
9430 automatically escaped as needed. Also, if file system or
9431 device paths are specified they are automatically turned
9432 into the appropriate mount or device unit names. Example:
9434 systemctl status /home
9435 systemctl status /dev/sda
9437 * The SysVConsole= configuration option has been removed from
9438 system.conf parsing.
9440 * The SysV search path is no longer exported on the D-Bus
9443 * The Names= option has been removed from unit file parsing.
9445 * There's a new man page bootup(7) detailing the boot process.
9447 * Every unit and every generator we ship with systemd now
9448 comes with full documentation. The self-explanatory boot is
9451 * A couple of services gained "systemd-" prefixes in their
9452 name if they wrap systemd code, rather than only external
9453 code. Among them fsck@.service which is now
9454 systemd-fsck@.service.
9456 * The HaveWatchdog property has been removed from the D-Bus
9459 * systemd.confirm_spawn= on the kernel command line should now
9462 * There's a new man page crypttab(5) which details all options
9463 we actually understand.
9465 * systemd-nspawn gained a new --capability= switch to pass
9466 additional capabilities to the container.
9468 * timedated will now read known NTP implementation unit names
9469 from /usr/lib/systemd/ntp-units.d/*.list,
9470 systemd-timedated-ntp.target has been removed.
9472 * journalctl gained a new switch "-b" that lists log data of
9473 the current boot only.
9475 * The notify socket is in the abstract namespace again, in
9476 order to support daemons which chroot() at start-up.
9478 * There is a new Storage= configuration option for journald
9479 which allows configuration of where log data should go. This
9480 also provides a way to disable journal logging entirely, so
9481 that data collected is only forwarded to the console, the
9482 kernel log buffer or another syslog implementation.
9484 * Many bugfixes and optimizations
9486 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
9487 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
9488 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
9489 Shawn Landden, Tom Gundersen
9493 * "systemctl help <unit>" now shows the man page if one is
9496 * Several new man pages have been added.
9498 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
9499 MaxLevelConsole= can now be specified in
9500 journald.conf. These options allow reducing the amount of
9501 data stored on disk or forwarded by the log level.
9503 * TimerSlackNSec= can now be specified in system.conf for
9504 PID1. This allows system-wide power savings.
9506 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
9507 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
9512 * logind is now capable of (optionally) handling power and
9513 sleep keys as well as the lid switch.
9515 * journalctl now understands the syntax "journalctl
9516 /usr/bin/avahi-daemon" to get all log output of a specific
9519 * CapabilityBoundingSet= in system.conf now also influences
9520 the capability bound set of usermode helpers of the kernel.
9522 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
9523 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
9524 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
9525 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
9529 * Note that we skipped 139 releases here in order to set the
9530 new version to something that is greater than both udev's
9531 and systemd's most recent version number.
9533 * udev: all udev sources are merged into the systemd source tree now.
9534 All future udev development will happen in the systemd tree. It
9535 is still fully supported to use the udev daemon and tools without
9536 systemd running, like in initramfs or other init systems. Building
9537 udev though, will require the *build* of the systemd tree, but
9538 udev can be properly *run* without systemd.
9540 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
9541 should be used to create dead device nodes as workarounds for broken
9544 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
9545 no longer supported. udev_monitor_new_from_netlink() needs to be
9546 used to subscribe to events.
9548 * udev: when udevd is started by systemd, processes which are left
9549 behind by forking them off of udev rules, are unconditionally cleaned
9550 up and killed now after the event handling has finished. Services or
9551 daemons must be started as systemd services. Services can be
9552 pulled-in by udev to get started, but they can no longer be directly
9553 forked by udev rules.
9555 * udev: the daemon binary is called systemd-udevd now and installed
9556 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
9557 to adapt to that, create symlink, or rename the binary after building
9560 * libudev no longer provides these symbols:
9561 udev_monitor_from_socket()
9562 udev_queue_get_failed_list_entry()
9563 udev_get_{dev,sys,run}_path()
9564 The versions number was bumped and symbol versioning introduced.
9566 * systemd-loginctl and systemd-journalctl have been renamed
9567 to loginctl and journalctl to match systemctl.
9569 * The config files: /etc/systemd/systemd-logind.conf and
9570 /etc/systemd/systemd-journald.conf have been renamed to
9571 logind.conf and journald.conf. Package updates should rename
9572 the files to the new names on upgrade.
9574 * For almost all files the license is now LGPL2.1+, changed
9575 from the previous GPL2.0+. Exceptions are some minor stuff
9576 of udev (which will be changed to LGPL2.1 eventually, too),
9577 and the MIT licensed sd-daemon.[ch] library that is suitable
9578 to be used as drop-in files.
9580 * systemd and logind now handle system sleep states, in
9581 particular suspending and hibernating.
9583 * logind now implements a sleep/shutdown/idle inhibiting logic
9584 suitable for a variety of uses. Soonishly Lennart will blog
9585 about this in more detail.
9587 * var-run.mount and var-lock.mount are no longer provided
9588 (which previously bind mounted these directories to their new
9589 places). Distributions which have not converted these
9590 directories to symlinks should consider stealing these files
9591 from git history and add them downstream.
9593 * We introduced the Documentation= field for units and added
9594 this to all our shipped units. This is useful to make it
9595 easier to explore the boot and the purpose of the various
9598 * All smaller setup units (such as
9599 systemd-vconsole-setup.service) now detect properly if they
9600 are run in a container and are skipped when
9601 appropriate. This guarantees an entirely noise-free boot in
9602 Linux container environments such as systemd-nspawn.
9604 * A framework for implementing offline system updates is now
9605 integrated, for details see:
9606 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
9608 * A new service type Type=idle is available now which helps us
9609 avoiding ugly interleaving of getty output and boot status
9612 * There's now a system-wide CapabilityBoundingSet= option to
9613 globally reduce the set of capabilities for the
9614 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
9615 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
9616 even CAP_NET_ADMIN system-wide for secure systems.
9618 * There are now system-wide DefaultLimitXXX= options to
9619 globally change the defaults of the various resource limits
9620 for all units started by PID 1.
9622 * Harald Hoyer's systemd test suite has been integrated into
9623 systemd which allows easy testing of systemd builds in qemu
9624 and nspawn. (This is really awesome! Ask us for details!)
9626 * The fstab parser is now implemented as generator, not inside
9629 * systemctl will now warn you if .mount units generated from
9630 /etc/fstab are out of date due to changes in fstab that
9631 have not been read by systemd yet.
9633 * systemd is now suitable for usage in initrds. Dracut has
9634 already been updated to make use of this. With this in place
9635 initrds get a slight bit faster but primarily are much
9636 easier to introspect and debug since "systemctl status" in
9637 the host system can be used to introspect initrd services,
9638 and the journal from the initrd is kept around too.
9640 * systemd-delta has been added, a tool to explore differences
9641 between user/admin configuration and vendor defaults.
9643 * PrivateTmp= now affects both /tmp and /var/tmp.
9645 * Boot time status messages are now much prettier and feature
9646 proper english language. Booting up systemd has never been
9649 * Read-ahead pack files now include the inode number of all
9650 files to pre-cache. When the inode changes the pre-caching
9651 is not attempted. This should be nicer to deal with updated
9652 packages which might result in changes of read-ahead
9655 * We now temporaritly lower the kernel's read_ahead_kb variable
9656 when collecting read-ahead data to ensure the kernel's
9657 built-in read-ahead does not add noise to our measurements
9658 of necessary blocks to pre-cache.
9660 * There's now RequiresMountsFor= to add automatic dependencies
9661 for all mounts necessary for a specific file system path.
9663 * MountAuto= and SwapAuto= have been removed from
9664 system.conf. Mounting file systems at boot has to take place
9667 * nspawn now learned a new switch --uuid= to set the machine
9668 ID on the command line.
9670 * nspawn now learned the -b switch to automatically search
9673 * vt102 is now the default TERM for serial TTYs, upgraded from
9676 * systemd-logind now works on VT-less systems.
9678 * The build tree has been reorganized. The individual
9679 components now have directories of their own.
9681 * A new condition type ConditionPathIsReadWrite= is now available.
9683 * nspawn learned the new -C switch to create cgroups for the
9684 container in other hierarchies.
9686 * We now have support for hardware watchdogs, configurable in
9689 * The scheduled shutdown logic now has a public API.
9691 * We now mount /tmp as tmpfs by default, but this can be
9692 masked and /etc/fstab can override it.
9694 * Since udisks does not make use of /media anymore we are not
9695 mounting a tmpfs on it anymore.
9697 * journalctl gained a new --local switch to only interleave
9698 locally generated journal files.
9700 * We can now load the IMA policy at boot automatically.
9702 * The GTK tools have been split off into a systemd-ui.
9704 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
9705 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
9706 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
9707 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
9708 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
9709 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
9710 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
9711 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
9712 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
9717 * This is mostly a bugfix release
9719 * Support optional initialization of the machine ID from the
9720 KVM or container configured UUID.
9722 * Support immediate reboots with "systemctl reboot -ff"
9724 * Show /etc/os-release data in systemd-analyze output
9726 * Many bugfixes for the journal, including endianness fixes and
9727 ensuring that disk space enforcement works
9729 * sd-login.h is C++ compatible again
9731 * Extend the /etc/os-release format on request of the Debian
9734 * We now refuse non-UTF8 strings used in various configuration
9735 and unit files. This is done to ensure we do not pass invalid
9736 data over D-Bus or expose it elsewhere.
9738 * Register Mimo USB Screens as suitable for automatic seat
9741 * Read SELinux client context from journal clients in a race
9744 * Reorder configuration file lookup order. /etc now always
9745 overrides /run in order to allow the administrator to always
9746 and unconditionally override vendor-supplied or
9747 automatically generated data.
9749 * The various user visible bits of the journal now have man
9750 pages. We still lack man pages for the journal API calls
9753 * We now ship all man pages in HTML format again in the
9756 Contributions from: Dave Reisner, Dirk Eibach, Frederic
9757 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
9758 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
9763 * This is mostly a bugfix release
9765 * systems lacking /etc/os-release are no longer supported.
9767 * Various functionality updates to libsystemd-login.so
9769 * Track class of PAM logins to distinguish greeters from
9772 Contributions from: Kay Sievers, Lennart Poettering, Michael
9777 * This is an important bugfix release for v41.
9779 * Building man pages is now optional which should be useful
9780 for those building systemd from git but unwilling to install
9783 * Watchdog support for supervising services is now usable. In
9784 a future release support for hardware watchdogs
9785 (i.e. /dev/watchdog) will be added building on this.
9787 * Service start rate limiting is now configurable and can be
9788 turned off per service. When a start rate limit is hit a
9789 reboot can automatically be triggered.
9791 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
9793 Contributions from: Benjamin Franzke, Bill Nottingham,
9794 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
9795 Schmidt, Michał Górny, Piotr Drąg
9799 * The systemd binary is installed /usr/lib/systemd/systemd now;
9800 An existing /sbin/init symlink needs to be adapted with the
9803 * The code that loads kernel modules has been ported to invoke
9804 libkmod directly, instead of modprobe. This means we do not
9805 support systems with module-init-tools anymore.
9807 * Watchdog support is now already useful, but still not
9810 * A new kernel command line option systemd.setenv= is
9811 understood to set system wide environment variables
9812 dynamically at boot.
9814 * We now limit the set of capabilities of systemd-journald.
9816 * We now set SIGPIPE to ignore by default, since it only is
9817 useful in shell pipelines, and has little use in general
9818 code. This can be disabled with IgnoreSIPIPE=no in unit
9821 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
9822 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
9827 * This is mostly a bugfix release
9829 * We now expose the reason why a service failed in the
9830 "Result" D-Bus property.
9832 * Rudimentary service watchdog support (will be completed over
9833 the next few releases.)
9835 * When systemd forks off in order execute some service we will
9836 now immediately changes its argv[0] to reflect which process
9837 it will execute. This is useful to minimize the time window
9838 with a generic argv[0], which makes bootcharts more useful
9840 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
9841 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
9842 Mike Kazantsev, Ray Strode
9846 * This is mostly a test release, but incorporates many
9849 * New systemd-cgtop tool to show control groups by their
9852 * Linking against libacl for ACLs is optional again. If
9853 disabled, support tracking device access for active logins
9854 goes becomes unavailable, and so does access to the user
9855 journals by the respective users.
9857 * If a group "adm" exists, journal files are automatically
9858 owned by them, thus allow members of this group full access
9859 to the system journal as well as all user journals.
9861 * The journal now stores the SELinux context of the logging
9862 client for all entries.
9864 * Add C++ inclusion guards to all public headers
9866 * New output mode "cat" in the journal to print only text
9867 messages, without any meta data like date or time.
9869 * Include tiny X server wrapper as a temporary stop-gap to
9870 teach XOrg udev display enumeration. This is used by display
9871 managers such as gdm, and will go away as soon as XOrg
9872 learned native udev hotplugging for display devices.
9874 * Add new systemd-cat tool for executing arbitrary programs
9875 with STDERR/STDOUT connected to the journal. Can also act as
9876 BSD logger replacement, and does so by default.
9878 * Optionally store all locally generated coredumps in the
9879 journal along with meta data.
9881 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
9882 writing short strings to files (for usage for /sys), and for
9883 creating symlinks, character and block device nodes.
9885 * New unit file option ControlGroupPersistent= to make cgroups
9886 persistent, following the mechanisms outlined in
9887 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
9889 * Support multiple local RTCs in a sane way
9891 * No longer monopolize IO when replaying readahead data on
9892 rotating disks, since we might starve non-file-system IO to
9893 death, since fanotify() will not see accesses done by blkid,
9896 * Do not show kernel threads in systemd-cgls anymore, unless
9897 requested with new -k switch.
9899 Contributions from: Dan Horák, Kay Sievers, Lennart
9900 Poettering, Michal Schmidt
9904 * This is mostly a test release, but incorporates many
9907 * The git repository moved to:
9908 git://anongit.freedesktop.org/systemd/systemd
9909 ssh://git.freedesktop.org/git/systemd/systemd
9911 * First release with the journal
9912 http://0pointer.de/blog/projects/the-journal.html
9914 * The journal replaces both systemd-kmsg-syslogd and
9915 systemd-stdout-bridge.
9917 * New sd_pid_get_unit() API call in libsystemd-logind
9919 * Many systemadm clean-ups
9921 * Introduce remote-fs-pre.target which is ordered before all
9922 remote mounts and may be used to start services before all
9925 * Added Mageia support
9927 * Add bash completion for systemd-loginctl
9929 * Actively monitor PID file creation for daemons which exit in
9930 the parent process before having finished writing the PID
9931 file in the daemon process. Daemons which do this need to be
9932 fixed (i.e. PID file creation must have finished before the
9933 parent exits), but we now react a bit more gracefully to them.
9935 * Add colourful boot output, mimicking the well-known output
9936 of existing distributions.
9938 * New option PassCredentials= for socket units, for
9939 compatibility with a recent kernel ABI breakage.
9941 * /etc/rc.local is now hooked in via a generator binary, and
9942 thus will no longer act as synchronization point during
9945 * systemctl list-unit-files now supports --root=.
9947 * systemd-tmpfiles now understands two new commands: z, Z for
9948 relabelling files according to the SELinux database. This is
9949 useful to apply SELinux labels to specific files in /sys,
9952 * Output of SysV services is now forwarded to both the console
9953 and the journal by default, not only just the console.
9955 * New man pages for all APIs from libsystemd-login.
9957 * The build tree got reorganized and the build system is a
9958 lot more modular allowing embedded setups to specifically
9959 select the components of systemd they are interested in.
9961 * Support for Linux systems lacking the kernel VT subsystem is
9964 * configure's --with-rootdir= got renamed to
9965 --with-rootprefix= to follow the naming used by udev and
9968 * Unless specified otherwise we will now install to /usr instead
9969 of /usr/local by default.
9971 * Processes with '@' in argv[0][0] are now excluded from the
9972 final shut-down killing spree, following the logic explained
9974 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
9976 * All processes remaining in a service cgroup when we enter
9977 the START or START_PRE states are now killed with
9978 SIGKILL. That means it is no longer possible to spawn
9979 background processes from ExecStart= lines (which was never
9980 supported anyway, and bad style).
9982 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
9983 reloading of units together.
9985 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
9986 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
9987 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
9988 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
9989 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
projects / platform / upstream / systemd.git / blob