1 systemd System and Service Manager
3 CHANGES WITH 244 in spe:
5 * Support for the cpuset cgroups v2 controller has been added.
6 Processes may be restricted to specific CPUs using the new
7 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
8 AllowedMemoryNodes= setting.
10 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
11 now be configured using a new RestartKillSignal= settting. This
12 allows units which signals to request termination to implement
13 different behaviour when stopping in preparation for a restart.
15 * "systemctl clean" may now be used also for socket, mount, and swap
18 * systemd will also read configuration options from the EFI variable
19 SystemdOptions. This may be used to configure systemd behaviour when
20 modifying the kernel command line is inconvenient, but configuration
21 on disk is read too late, for example for the options related to
22 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
25 * systemd will now disable printk ratelimits in early boot. This should
26 allow us to capture more logs from the early boot phase where normal
27 storage is not available and the kernel ring buffer is used for
28 logging. Configuration on the kernel command line has higher priority
29 and overrides the systemd setting.
31 systemd programs which log to /dev/kmsg directly use internal
32 ratelimits to prevent runaway logging. (Normally this is only used
33 during early boot, so in practice this change has very little
36 * Unit files now support top level dropin directories of the form
37 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
38 that affects all corresponding unit files.
40 * systemctl gained support for 'stop --job-mode=triggering' which will
41 stop the specified unit and any units which could trigger it.
43 * Unit status display now includes units triggering and triggered by
46 * The RuntimeMaxSec= setting is now supported by scopes, not just
47 .service units. This is particularly useful for PAM sessions which
48 create a scope unit for the user login. systemd.runtime_max_sec=
49 setting may used with the pam_systemd module to limit the duration
50 of the PAM session, for example for time-limited logins.
52 * A new @pkey system call group is now defined to make it easier to
53 whitelist memory protection syscalls for containers and services
54 which need to use them.
56 * systemd-udevd: removed the 30s timeout for killing stale workers on
57 exit. systemd-udevd now waits for workers to finish. The hard-coded
58 exit timeout of 30s was too short for some large installations, where
59 driver initialization could be prematurely interrupted during initrd
60 processing if the root file system had been mounted and init was
61 preparing to switch root. If udevd is run without systemd and workers
62 are hanging while udevd receives an exit signal, udevd will now exit
63 when udev.event_timeout is reached for the last hanging worker. With
64 systemd, the exit timeout can additionally be configured using
65 TimeoutStopSec= in systemd-udevd.service.
67 * udev now provides a program (fido_id) that identifies FIDO CTAP1
68 ("U2F")/CTAP2 security tokens based on the usage declared in their
69 report and descriptor and outputs suitable environment variables.
70 This replaces the externally maintained whitelists of all known
71 security tokens that were used previously.
73 * Automatically generated autosuspend udev rules for whitelisted
74 devices have been imported from the Chromium OS project. This should
75 improve power saving with many more devices.
77 * udev gained a new "CONST{key}=value" setting that allows matching
78 against system-wide constants without forking a helper binary.
79 Currently "arch" and "virt" keys are supported.
81 * udev now opens CDROMs in non-exclusive mode when querying their
82 capabilities. This should fix issues where other programs trying to
83 use the CDROM cannot gain access to it, but carries a risk of
84 interfering with programs writing to the disk, if they did not open
85 the device in exclusive mode as they should.
87 * systemd-networkd does not create a default route for IPv4 link local
88 addressing anymore. The creation of the route was unexpected and was
89 breaking routing in various cases, but people who rely on it being
90 created implicitly will need to adjust. Such a route may be requested
91 with DefaultRouteOnDevice=yes.
93 Similarly, systemd-networkd will not assign a link-local IPv6 address
94 when IPv6 link-local routing is not enabled.
96 * Receive and transmit buffers may now be configured on links with
97 the new RxBufferSize= and TxBufferSize= settings.
99 * systemd-networkd may now advertise additional IPv6 routes. A new
100 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
103 * systemd-networkd may now configure "next hop" routes using the
104 [NextHop] section and Gateway= and Id= settings.
106 * systemd-networkd will now retain DHCP config on restarts by default
107 (but this may be overridden using the KeepConfiguration= setting).
108 The default for SendRelease= has been changed to true.
110 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
111 received from the server.
113 The client will use the received SIP server list if UseSIP=yes is
116 The client may be configured to request specific options from the
117 server using a new RequestOptions= setting.
119 The client may be configured to send arbitrary options to the server
120 using a new SendOption= setting.
122 A new IPServiceType= setting has been added to configure the "IP
123 service type" value used by the client.
125 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
126 request prefix hints in the DHCPv6 solicitation.
128 * The DHCPv4 server may be configured to send arbitrary options using
129 a new SendOption= setting.
131 * The DHCPv4 server may now be configured to emit SIP server list using
132 the new EmitSIP= and SIP= settings.
134 * systemd-networkd and networkctl may now renew DHCP leases on demand.
135 networkctl has a new 'networkctl renew' verb.
137 * systemd-networkd may now reconfigure links on demand. networkctl
138 gained two new verbs: "reload" will reload the configuration, and
139 "reconfigure DEVICE…" will reconfigure one or more devices.
141 * .network files may now match on SSID and BSSID of a wireless network,
142 i.e. the access point name and hardware address using the new SSID=
143 and BSSID= options. networkctl will display the current SSID and
144 BSSID for wireless links.
146 .network files may also match on the wireless network type using the
147 new WLANInterfaceType= option.
149 * systemd-networkd now includes default configuration that enables
150 link-local addressing when connected to an ad-hoc wireless network.
152 * systemd-networkd may configure the Traffic Control queueing
153 disciplines in the kernel using the new
154 [TrafficControlQueueingDiscipline] section and Parent=,
155 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
156 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
157 NetworkEmulatorDuplicateRate= settings.
159 * systemd-tmpfiles gained a new w+ setting to append to files.
161 * systemd-analyze dump will now report when the memory configuration in
162 the kernel does not match what systemd has configured (usually,
163 because some external program has modified the kernel configuration
166 * systemd-analyze gained a new --base-time= switch instructs the
167 'calendar' verb to resolve times relative to that timestamp instead
170 * journalctl --update-catalog now produces deterministic output (making
171 reproducible image builds easier).
173 * A new devicetree-overlay setting is now documented in the Boot Loader
176 * The default value of the WatchdogSec= setting used in systemd
177 services (the ones bundled with the project itself) may be set at
178 configuration time using the -Dservice-watchdog= setting. If set to
179 empty, the watchdogs will be disabled.
181 * systemd-resolved validates IP addresses in certificates now when GnuTLS
184 * libcryptsetup >= 2.0.1 is now required.
186 * A configuration option -Duser-path= may be used to override the $PATH
187 used by the user service manager. The default is again to use the same
188 path as the system manager.
192 * This release enables unprivileged programs (i.e. requiring neither
193 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
194 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
195 kernel for the whole UNIX group range, i.e. all processes. This
196 change should be reasonably safe, as the kernel support for it was
197 specifically implemented to allow safe access to ICMP Echo for
198 processes lacking any privileges. If this is not desirable, it can be
199 disabled again by setting the parameter to "1 0".
201 * Previously, filters defined with SystemCallFilter= would have the
202 effect that any calling of an offending system call would terminate
203 the calling thread. This behaviour never made much sense, since
204 killing individual threads of unsuspecting processes is likely to
205 create more problems than it solves. With this release the default
206 action changed from killing the thread to killing the whole
207 process. For this to work correctly both a kernel version (>= 4.14)
208 and a libseccomp version (>= 2.4.0) supporting this new seccomp
209 action is required. If an older kernel or libseccomp is used the old
210 behaviour continues to be used. This change does not affect any
211 services that have no system call filters defined, or that use
212 SystemCallErrorNumber= (and thus see EPERM or another error instead
213 of being killed when calling an offending system call). Note that
214 systemd documentation always claimed that the whole process is
215 killed. With this change behaviour is thus adjusted to match the
218 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
219 4194304 by default, i.e. the full 22bit range the kernel allows, up
220 from the old 16bit range. This should improve security and
221 robustness, as PID collisions are made less likely (though certainly
222 still possible). There are rumours this might create compatibility
223 problems, though at this moment no practical ones are known to
224 us. Downstream distributions are hence advised to undo this change in
225 their builds if they are concerned about maximum compatibility, but
226 for everybody else we recommend leaving the value bumped. Besides
227 improving security and robustness this should also simplify things as
228 the maximum number of allowed concurrent tasks was previously bounded
229 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
230 only a single knob is left ("kernel.threads-max"). There have been
231 concerns that usability is affected by this change because larger PID
232 numbers are harder to type, but we believe the change from 5 digits
233 to 7 digits doesn't hamper usability.
235 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
236 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
237 hierarchically set default memory protection values for a particular
238 subtree of the unit hierarchy.
240 * Memory protection directives can now take a value of zero, allowing
241 explicit opting out of a default value propagated by an ancestor.
243 * systemd now defaults to the "unified" cgroup hierarchy setup during
244 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
245 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
246 change reflects the fact that cgroupsv2 support has matured
247 substantially in both systemd and in the kernel, and is clearly the
248 way forward. Downstream production distributions might want to
249 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
250 their builds as unfortunately the popular container managers have not
251 caught up with the kernel API changes.
253 * Man pages are not built by default anymore (html pages were already
254 disabled by default), to make development builds quicker. When
255 building systemd for a full installation with documentation, meson
256 should be called with -Dman=true and/or -Dhtml=true as appropriate.
257 The default was changed based on the assumption that quick one-off or
258 repeated development builds are much more common than full optimized
259 builds for installation, and people need to pass various other
260 options to when doing "proper" builds anyway, so the gain from making
261 development builds quicker is bigger than the one time disruption for
264 Two scripts are created in the *build* directory to generate and
265 preview man and html pages on demand, e.g.:
267 build/man/man systemctl
268 build/man/html systemd.index
270 * libidn2 is used by default if both libidn2 and libidn are installed.
271 Please use -Dlibidn=true if libidn is preferred.
273 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
274 big-endian machines. Before, bytes were written and read in native
275 machine order as exposed by the native libc __cpu_mask interface.
276 Now, little-endian order is always used (CPUs 0–7 are described by
277 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
278 This change fixes D-Bus calls that cross endianness boundary.
280 The presentation format used for CPUAffinity= by "systemctl show" and
281 "systemd-analyze dump" is changed to present CPU indices instead of
282 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
283 shown as CPUAffinity=03000000000000000000000000000… (on
284 little-endian) or CPUAffinity=00000000000000300000000000000… (on
285 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
286 input format. The maximum integer that will be printed in the new
287 format is 8191 (four digits), while the old format always used a very
288 long number (with the length varying by architecture), so they can be
289 unambiguously distinguished.
291 * /usr/sbin/halt.local is no longer supported. Implementation in
292 distributions was inconsistent and it seems this functionality was
295 To replace this functionality, users should:
296 - either define a new unit and make it a dependency of final.target
297 (systemctl add-wants final.target my-halt-local.service)
298 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
299 and ensure that it accepts "halt", "poweroff", "reboot", and
300 "kexec" as an argument, see the description in systemd-shutdown(8).
302 * When a [Match] section in .link or .network file is empty (contains
303 no match patterns), a warning will be emitted. Please add any "match
304 all" pattern instead, e.g. OriginalName=* or Name=* in case all
305 interfaces should really be matched.
307 * A new setting NUMAPolicy= may be used to set process memory
308 allocation policy. This setting can be specified in
309 /etc/systemd/system.conf and hence will set the default policy for
310 PID1. The default policy can be overridden on a per-service
311 basis. The related setting NUMAMask= is used to specify NUMA node
312 mask that should be associated with the selected policy.
314 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
315 generates when processes it manages are reaching their memory limits,
316 and will place their units in a special state, and optionally kill or
319 * The service manager will now expose bus properties for the IO
320 resources used by units. This information is also shown in "systemctl
321 status" now (for services that have IOAccounting=yes set). Moreover,
322 the IO accounting data is included in the resource log message
323 generated whenever a unit stops.
325 * Units may now configure an explicit time-out to wait for when killed
326 with SIGABRT, for example when a service watchdog is hit. Previously,
327 the regular TimeoutStopSec= time-out was applied in this case too —
328 now a separate time-out may be set using TimeoutAbortSec=.
330 * Services may now send a special WATCHDOG=trigger message with
331 sd_notify() to trigger an immediate "watchdog missed" event, and thus
332 trigger service termination. This is useful both for testing watchdog
333 handling, but also for defining error paths in services, that shall
334 be handled the same way as watchdog events.
336 * There are two new per-unit settings IPIngressFilterPath= and
337 IPEgressFilterPath= which allow configuration of a BPF program
338 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
339 to apply to the IP packet ingress/egress path of all processes of a
340 unit. This is useful to allow running systemd services with BPF
341 programs set up externally.
343 * systemctl gained a new "clean" verb for removing the state, cache,
344 runtime or logs directories of a service while it is terminated. The
345 new verb may also be used to remove the state maintained on disk for
346 timer units that have Persistent= configured.
348 * During the last phase of shutdown systemd will now automatically
349 increase the log level configured in the "kernel.printk" sysctl so
350 that any relevant loggable events happening during late shutdown are
351 made visible. Previously, loggable events happening so late during
352 shutdown were generally lost if the "kernel.printk" sysctl was set to
353 high thresholds, as regular logging daemons are terminated at that
354 time and thus nothing is written to disk.
356 * If processes terminated during the last phase of shutdown do not exit
357 quickly systemd will now show their names after a short time, to make
358 debugging easier. After a longer time-out they are forcibly killed,
361 * journalctl (and the other tools that display logs) will now highlight
362 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
363 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
364 are now shown in blue color, to separate them visually from regular
365 logs. References to configuration files are now turned into clickable
366 links on terminals that support that.
368 * systemd-journald will now stop logging to /var/log/journal during
369 shutdown when /var/ is on a separate mount, so that it can be
370 unmounted safely during shutdown.
372 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
374 * systemd-resolved "Cache=" configuration option in resolved.conf has
375 been extended to also accept the 'no-negative' value. Previously,
376 only a boolean option was allowed (yes/no), having yes as the
377 default. If this option is set to 'no-negative', negative answers are
378 not cached while the old cache heuristics are used positive answers.
379 The default remains unchanged.
381 * The predictable naming scheme for network devices now supports
382 generating predictable names for "netdevsim" devices.
384 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
387 Those two changes form a new net.naming-policy-scheme= entry.
388 Distributions which want to preserve naming stability may want to set
389 the -Ddefault-net-naming-scheme= configuration option.
391 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
394 * systemd-networkd's bridge FDB support now allows configuration of a
395 destination address for each entry (Destination=), as well as the
396 VXLAN VNI (VNI=), as well as an option to declare what an entry is
397 associated with (AssociatedWith=).
399 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
400 option for configuring the maximum number of DHCP lease requests. It
401 also learnt a new BlackList= option for blacklisting DHCP servers (a
402 similar setting has also been added to the IPv6 RA client), as well
403 as a SendRelease= option for configuring whether to send a DHCP
404 RELEASE message when terminating.
406 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
407 separately in the [DHCPv4] and [DHCPv6] sections.
409 * systemd-networkd's DHCP support will now optionally create an
410 implicit host route to the DNS server specified in the DHCP lease, in
411 addition to the routes listed explicitly in the lease. This should
412 ensure that in multi-homed systems DNS traffic leaves the systems on
413 the interface that acquired the DNS server information even if other
414 routes such as default routes exist. This behaviour may be turned on
415 with the new RoutesToDNS= option.
417 * systemd-networkd's VXLAN support gained a new option
418 GenericProtocolExtension= for enabling VXLAN Generic Protocol
419 Extension support, as well as IPDoNotFragment= for setting the IP
420 "Don't fragment" bit on outgoing packets. A similar option has been
421 added to the GENEVE support.
423 * In systemd-networkd's [Route] section you may now configure
424 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
425 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
426 propagation. The Type= setting now supports local, broadcast,
427 anycast, multicast, any, xresolve routes, too.
429 * systemd-networkd's [Network] section learnt a new option
430 DefaultRouteOnDevice= for automatically configuring a default route
431 onto the network device.
433 * systemd-networkd's bridging support gained two new options ProxyARP=
434 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
435 MulticastRouter= for configuring multicast routing behaviour. A new
436 option MulticastIGMPVersion= may be used to change bridge's multicast
437 Internet Group Management Protocol (IGMP) version.
439 * systemd-networkd's FooOverUDP support gained the ability to configure
440 local and peer IP addresses via Local= and Peer=. A new option
441 PeerPort= may be used to configure the peer's IP port.
443 * systemd-networkd's TUN support gained a new setting VnetHeader= for
444 tweaking Generic Segment Offload support.
446 * networkctl gained a new "delete" command for removing virtual network
447 devices, as well as a new "--stats" switch for showing device
450 * networkd.conf gained a new setting SpeedMeter= and
451 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
452 measured speed may be shown by 'networkctl status'.
454 * "networkctl status" now displays MTU and queue lengths, and more
455 detailed information about VXLAN and bridge devices.
457 * systemd-networkd's .network and .link files gained a new Property=
458 setting in the [Match] section, to match against devices with
459 specific udev properties.
461 * systemd-networkd's tunnel support gained a new option
462 AssignToLoopback= for selecting whether to use the loopback device
463 "lo" as underlying device.
465 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
466 been renamed to LinkLayerAddress=, and it now allows configuration of
469 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
470 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
471 IPv6 configuration (static or DHCPv6) was found for a given
472 interface. It will not touch the sysctl otherwise.
474 * The order of entries is $PATH used by the user manager instance was
475 changed to put bin/ entries before the corresponding sbin/ entries.
476 It is recommended to not rely on this order, and only ever have one
477 binary with a given name in the system paths under /usr.
479 * A new tool systemd-network-generator has been added that may generate
480 .network, .netdev and .link files from IP configuration specified on
481 the kernel command line in the format used by Dracut.
483 * The CriticalConnection= setting in .network files is now deprecated,
484 and replaced by a new KeepConfiguration= setting which allows more
485 detailed configuration of the IP configuration to keep in place.
487 * systemd-analyze gained a few new verbs:
489 - "systemd-analyze timestamp" parses and converts timestamps. This is
490 similar to the existing "systemd-analyze calendar" command which
491 does the same for recurring calendar events.
493 - "systemd-analyze timespan" parses and converts timespans (i.e.
494 durations as opposed to points in time).
496 - "systemd-analyze condition" will parse and test ConditionXYZ=
499 - "systemd-analyze exit-status" will parse and convert exit status
500 codes to their names and back.
502 - "systemd-analyze unit-files" will print a list of all unit
503 file paths and unit aliases.
505 * SuccessExitStatus=, RestartPreventExitStatus=, and
506 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
507 is equivalent to "65"). Those exit status name mappings may be
508 displayed with the systemd-analyze exit-status verb describe above.
510 * systemd-logind now exposes a per-session SetBrightness() bus call,
511 which may be used to securely change the brightness of a kernel
512 brightness device, if it belongs to the session's seat. By using this
513 call unprivileged clients can make changes to "backlight" and "leds"
514 devices securely with strict requirements on session membership.
515 Desktop environments may use this to generically make brightness
516 changes to such devices without shipping private SUID binaries or
517 udev rules for that purpose.
519 * "udevadm info" gained a --wait-for-initialization switch to wait for
520 a device to be initialized.
522 * systemd-hibernate-resume-generator will now look for resumeflags= on
523 the kernel command line, which is similar to rootflags= and may be
524 used to configure device timeout for the hibernation device.
526 * sd-event learnt a new API call sd_event_source_disable_unref() for
527 disabling and unref'ing an event source in a single function. A
528 related call sd_event_source_disable_unrefp() has been added for use
529 with gcc's cleanup extension.
531 * The sd-id128.h public API gained a new definition
532 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
535 * "busctl introspect" gained a new switch --xml-interface for dumping
536 XML introspection data unmodified.
538 * PID 1 may now show the unit name instead of the unit description
539 string in its status output during boot. This may be configured in
540 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
541 kernel command line option systemd.status_unit_format=.
543 * PID 1 now understands a new option KExecWatchdogSec= in
544 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
545 Previously watchdog functionality was only available for regular
546 reboots. The new setting defaults to off, because we don't know in
547 the general case if the watchdog will be reset after kexec (some
548 drivers do reset it, but not all), and the new userspace might not be
549 configured to handle the watchdog.
551 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
552 RebootWatchdogSec= to more clearly communicate what it is about. The
553 old name is still accepted for compatibility.
555 * The systemd.debug_shell kernel command line option now optionally
556 takes a tty name to spawn the debug shell on, which allows a
557 different tty to be selected than the built-in default.
559 * Service units gained a new ExecCondition= setting which will run
560 before ExecStartPre= and either continue execution of the unit (for
561 clean exit codes), stop execution without marking the unit failed
562 (for exit codes 1 through 254), or stop execution and fail the unit
563 (for exit code 255 or abnormal termination).
565 * A new service systemd-pstore.service has been added that pulls data
566 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
569 * timedatectl gained new verbs for configuring per-interface NTP
570 service configuration for systemd-timesyncd.
572 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
573 2019. (You can set non-UTF-8 locales though, if you know their name.)
575 * If variable assignments in sysctl.d/ files are prefixed with "-" any
576 failures to apply them are now ignored.
578 * systemd-random-seed.service now optionally credits entropy when
579 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
580 true for the service to enable this behaviour, but please consult the
581 documentation first, since this comes with a couple of caveats.
583 * systemd-random-seed.service is now a synchronization point for full
584 initialization of the kernel's entropy pool. Services that require
585 /dev/urandom to be correctly initialized should be ordered after this
588 * The systemd-boot boot loader has been updated to optionally maintain
589 a random seed file in the EFI System Partition (ESP). During the boot
590 phase, this random seed is read and updated with a new seed
591 cryptographically derived from it. Another derived seed is passed to
592 the OS. The latter seed is then credited to the kernel's entropy pool
593 very early during userspace initialization (from PID 1). This allows
594 systems to boot up with a fully initialized kernel entropy pool from
595 earliest boot on, and thus entirely removes all entropy pool
596 initialization delays from systems using systemd-boot. Special care
597 is taken to ensure different seeds are derived on system images
598 replicated to multiple systems. "bootctl status" will show whether
599 a seed was received from the boot loader.
601 * bootctl gained two new verbs:
603 - "bootctl random-seed" will generate the file in ESP and an EFI
604 variable to allow a random seed to be passed to the OS as described
607 - "bootctl is-installed" checks whether systemd-boot is currently
610 * bootctl will warn if it detects that boot entries are misconfigured
611 (for example if the kernel image was removed without purging the
614 * A new document has been added describing systemd's use and support
615 for the kernel's entropy pool subsystem:
617 https://systemd.io/RANDOM_SEEDS
619 * When the system is hibernated the swap device to write the
620 hibernation image to is now automatically picked from all available
621 swap devices, preferring the swap device with the highest configured
622 priority over all others, and picking the device with the most free
623 space if there are multiple devices with the highest priority.
625 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
626 option that permits selecting the timout how long to wait for a
627 device with an encryption key before asking for the password.
629 * IOWeight= has learnt to properly set the IO weight when using the
630 BFQ scheduler officially found in kernels 5.0+.
632 * A new mailing list has been created for reporting of security issues:
633 systemd-security@redhat.com. For mode details, see
634 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
636 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
637 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
638 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
639 Chiu, Chris Down, Christian Kellner, Clinton Roy, Connor Reeder, Daniel
640 Black, Daniele Medri, Dan Streetman, Dave Reisner, Dave Ross, David
641 Art, David Tardon, Debarshi Ray, Dimitri John Ledkov, Dominick Grift,
642 Donald Buczek, Douglas Christman, Eric DeVolder, EtherGraf, Evgeny
643 Vereshchagin, Feldwor, Felix Riemann, Florian Dollinger, Francesco
644 Pennica, Franck Bui, Frantisek Sumsal, Franz Pletz, frederik, Hans
645 de Goede, Iago López Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer,
646 Jack, Jakob Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan
647 Pokorný, Jan Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller,
648 Jérémy Rosen, Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann
649 B. Guðmundsson, Johannes Christ, Johannes Schmitz, Jonathan Rouleau,
650 Jorge Niedbalski, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
651 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
652 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
653 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
654 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
655 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
656 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
657 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Roberto
658 Santalla, Ronan Pigott, root, RussianNeuroMancer, Sebastian Jennen,
659 shinygold, Shreyas Behera, Simon Schricker, Susant Sahani, Thadeu Lima
660 de Souza Cascardo, Theo Ouzhinski, Thiebaud Weksteen, Thomas Haller,
661 Thomas Weißschuh, Tomas Mraz, Tommi Rantala, Topi Miettinen, VD-Lycos,
662 ven, Wieland Hoffmann, William A. Kennington III, William Wold, Xi
663 Ruoyao, Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew
664 Jędrzejewski-Szmek, Zhang Xianwei
666 – Camerino, 2019-09-03
670 * In .link files, MACAddressPolicy=persistent (the default) is changed
671 to cover more devices. For devices like bridges, tun, tap, bond, and
672 similar interfaces that do not have other identifying information,
673 the interface name is used as the basis for persistent seed for MAC
674 and IPv4LL addresses. The way that devices that were handled
675 previously is not changed, and this change is about covering more
676 devices then previously by the "persistent" policy.
678 MACAddressPolicy=random may be used to force randomized MACs and
679 IPv4LL addresses for a device if desired.
681 Hint: the log output from udev (at debug level) was enhanced to
682 clarify what policy is followed and which attributes are used.
683 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
684 may be used to view this.
686 Hint: if a bridge interface is created without any slaves, and gains
687 a slave later, then now the bridge does not inherit slave's MAC.
688 To inherit slave's MAC, for example, create the following file:
690 # /etc/systemd/network/98-bridge-inherit-mac.link
695 MACAddressPolicy=none
698 * The .device units generated by systemd-fstab-generator and other
699 generators do not automatically pull in the corresponding .mount unit
700 as a Wants= dependency. This means that simply plugging in the device
701 will not cause the mount unit to be started automatically. But please
702 note that the mount unit may be started for other reasons, in
703 particular if it is part of local-fs.target, and any unit which
704 (transitively) depends on local-fs.target is started.
706 * networkctl list/status/lldp now accept globbing wildcards for network
707 interface names to match against all existing interfaces.
709 * The $PIDFILE environment variable is set to point the absolute path
710 configured with PIDFile= for processes of that service.
712 * The fallback DNS server list was augmented with Cloudflare public DNS
713 servers. Use `-Ddns-servers=` to set a different fallback.
715 * A new special target usb-gadget.target will be started automatically
716 when a USB Device Controller is detected (which means that the system
717 is a USB peripheral).
719 * A new unit setting CPUQuotaPeriodSec= assigns the time period
720 relatively to which the CPU time quota specified by CPUQuota= is
723 * A new unit setting ProtectHostname= may be used to prevent services
724 from modifying hostname information (even if they otherwise would
725 have privileges to do so).
727 * A new unit setting NetworkNamespacePath= may be used to specify a
728 namespace for service or socket units through a path referring to a
729 Linux network namespace pseudo-file.
731 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
732 have an effect on .socket units: when used the listening socket is
733 created within the configured network namespace instead of the host
736 * ExecStart= command lines in unit files may now be prefixed with ':'
737 in which case environment variable substitution is
738 disabled. (Supported for the other ExecXYZ= settings, too.)
740 * .timer units gained two new boolean settings OnClockChange= and
741 OnTimezoneChange= which may be used to also trigger a unit when the
742 system clock is changed or the local timezone is
743 modified. systemd-run has been updated to make these options easily
744 accessible from the command line for transient timers.
746 * Two new conditions for units have been added: ConditionMemory= may be
747 used to conditionalize a unit based on installed system
748 RAM. ConditionCPUs= may be used to conditionalize a unit based on
751 * The @default system call filter group understood by SystemCallFilter=
752 has been updated to include the new rseq() system call introduced in
755 * A new time-set.target has been added that indicates that the system
756 time has been set from a local source (possibly imprecise). The
757 existing time-sync.target is stronger and indicates that the time has
758 been synchronized with a precise external source. Services where
759 approximate time is sufficient should use the new target.
761 * "systemctl start" (and related commands) learnt a new
762 --show-transaction option. If specified brief information about all
763 jobs queued because of the requested operation is shown.
765 * systemd-networkd recognizes a new operation state 'enslaved', used
766 (instead of 'degraded' or 'carrier') for interfaces which form a
767 bridge, bond, or similar, and an new 'degraded-carrier' operational
768 state used for the bond or bridge master interface when one of the
769 enslaved devices is not operational.
771 * .network files learnt the new IgnoreCarrierLoss= option for leaving
772 networks configured even if the carrier is lost.
774 * The RequiredForOnline= setting in .network files may now specify a
775 minimum operational state required for the interface to be considered
776 "online" by systemd-networkd-wait-online. Related to this
777 systemd-networkd-wait-online gained a new option --operational-state=
778 to configure the same, and its --interface= option was updated to
779 optionally also take an operational state specific for an interface.
781 * systemd-networkd-wait-online gained a new setting --any for waiting
782 for only one of the requested interfaces instead of all of them.
784 * systemd-networkd now implements L2TP tunnels.
786 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
787 may be used to cause autonomous and onlink prefixes received in IPv6
788 Router Advertisements to be ignored.
790 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
791 file settings may be used to tweak bridge behaviour.
793 * The new TripleSampling= option in .network files may be used to
794 configure CAN triple sampling.
796 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
797 used to point to private or preshared key for a WireGuard interface.
799 * /etc/crypttab now supports the same-cpu-crypt and
800 submit-from-crypt-cpus options to tweak encryption work scheduling
803 * systemd-tmpfiles will now take a BSD file lock before operating on a
804 contents of directory. This may be used to temporarily exclude
805 directories from aging by taking the same lock (useful for example
806 when extracting a tarball into /tmp or /var/tmp as a privileged user,
807 which might create files with really old timestamps, which
808 nevertheless should not be deleted). For further details, see:
810 https://systemd.io/TEMPORARY_DIRECTORIES
812 * systemd-tmpfiles' h line type gained support for the
813 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
814 controlling project quota inheritance.
816 * sd-boot and bootctl now implement support for an Extended Boot Loader
817 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
818 addition to the ESP partition mounted to /efi or /boot/efi.
819 Configuration file fragments, kernels, initrds and other EFI images
820 to boot will be loaded from both the ESP and XBOOTLDR partitions.
821 The XBOOTLDR partition was previously described by the Boot Loader
822 Specification, but implementation was missing in sd-boot. Support for
823 this concept allows using the sd-boot boot loader in more
824 conservative scenarios where the boot loader itself is placed in the
825 ESP but the kernels to boot (and their metadata) in a separate
828 * A system may now be booted with systemd.volatile=overlay on the
829 kernel command line, which causes the root file system to be set up
830 an overlayfs mount combining the root-only root directory with a
831 writable tmpfs. In this setup, the underlying root device is not
832 modified, and any changes are lost at reboot.
834 * Similar, systemd-nspawn can now boot containers with a volatile
835 overlayfs root with the new --volatile=overlay switch.
837 * systemd-nspawn can now consume OCI runtime bundles using a new
838 --oci-bundle= option. This implementation is fully usable, with most
839 features in the specification implemented, but since this a lot of
840 new code and functionality, this feature should most likely not
841 be used in production yet.
843 * systemd-nspawn now supports various options described by the OCI
844 runtime specification on the command-line and in .nspawn files:
845 --inaccessible=/Inaccessible= may be used to mask parts of the file
846 system tree, --console=/--pipe may be used to configure how standard
847 input, output, and error are set up.
849 * busctl learned the `emit` verb to generate D-Bus signals.
851 * systemd-analyze cat-config may be used to gather and display
852 configuration spread over multiple files, for example system and user
853 presets, tmpfiles.d, sysusers.d, udev rules, etc.
855 * systemd-analyze calendar now takes an optional new parameter
856 --iterations= which may be used to show a maximum number of iterations
857 the specified expression will elapse next.
859 * The sd-bus C API gained support for naming method parameters in the
862 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
863 the reboot() system call expects.
865 * journalctl learnt a new --cursor-file= option that points to a file
866 from which a cursor should be loaded in the beginning and to which
867 the updated cursor should be stored at the end.
869 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
870 detected by systemd-detect-virt (and may also be used in
871 ConditionVirtualization=).
873 * The behaviour of systemd-logind may now be modified with environment
874 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
875 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
876 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
877 skip the relevant operation completely (when set to false), or to
878 create a flag file in /run/systemd (when set to true), instead of
879 actually commencing the real operation when requested. The presence
880 of /run/systemd/reboot-to-firmware-setup,
881 /run/systemd/reboot-to-boot-loader-menu, and
882 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
883 boot loader implementations to replace some steps logind performs
884 during reboot with their own operations.
886 * systemctl can be used to request a reboot into the boot loader menu
887 or a specific boot loader entry with the new --boot-load-menu= and
888 --boot-loader-entry= options to a reboot command. (This requires a
889 boot loader that supports this, for example sd-boot.)
891 * kernel-install will no longer unconditionally create the output
892 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
893 snippets, but will do only if the machine-specific parent directory
894 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
895 to create this parent directory during sd-boot installation.
897 This makes it easier to use kernel-install with plugins which support
898 a different layout of the bootloader partitions (for example grub2).
900 * During package installation (with `ninja install`), we would create
901 symlinks for getty@tty1.service, systemd-networkd.service,
902 systemd-networkd.socket, systemd-resolved.service,
903 remote-cryptsetup.target, remote-fs.target,
904 systemd-networkd-wait-online.service, and systemd-timesyncd.service
905 in /etc, as if `systemctl enable` was called for those units, to make
906 the system usable immediately after installation. Now this is not
907 done anymore, and instead calling `systemctl preset-all` is
908 recommended after the first installation of systemd.
910 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
911 is built on seccomp. When turned on creation of SUID/SGID files is
914 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
915 implied if DynamicUser= is turned on for a service. This hardens
916 these services, so that they neither can benefit from nor create
917 SUID/SGID executables. This is a minor compatibility breakage, given
918 that when DynamicUser= was first introduced SUID/SGID behaviour was
919 unaffected. However, the security benefit of these two options is
920 substantial, and the setting is still relatively new, hence we opted
921 to make it mandatory for services with dynamic users.
923 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
924 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
925 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
926 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
927 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
928 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
929 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
930 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
931 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
932 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
933 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
934 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
935 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
936 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
937 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
938 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
939 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
940 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
946 * The default locale can now be configured at compile time. Otherwise,
947 a suitable default will be selected automatically (one of C.UTF-8,
950 * The version string shown by systemd and other tools now includes the
951 git commit hash when built from git. An override may be specified
952 during compilation, which is intended to be used by distributions to
953 include the package release information.
955 * systemd-cat can now filter standard input and standard error streams
956 for different syslog priorities using the new --stderr-priority=
959 * systemd-journald and systemd-journal-remote reject entries which
960 contain too many fields (CVE-2018-16865) and set limits on the
961 process' command line length (CVE-2018-16864).
963 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
966 * A new network device NamePolicy "keep" is implemented for link files,
967 and used by default in 99-default.link (the fallback configuration
968 provided by systemd). With this policy, if the network device name
969 was already set by userspace, the device will not be renamed again.
970 This matches the naming scheme that was implemented before
971 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
972 is also enabled by default, even if not specified. Effectively, this
973 means that if naming-scheme >= 240 is specified, network devices will
974 be renamed according to the configuration, even if they have been
975 renamed already, if "keep" is not specified as the naming policy in
976 the .link file. The 99-default.link file provided by systemd includes
977 "keep" for backwards compatibility, but it is recommended for user
978 installed .link files to *not* include it.
980 The "kernel" policy, which keeps kernel names declared to be
981 "persistent", now works again as documented.
983 * kernel-install script now optionally takes the paths to one or more
984 initrd files, and passes them to all plugins.
986 * The mincore() system call has been dropped from the @system-service
987 system call filter group, as it is pretty exotic and may potentially
988 used for side-channel attacks.
990 * -fPIE is dropped from compiler and linker options. Please specify
991 -Db_pie=true option to meson to build position-independent
992 executables. Note that the meson option is supported since meson-0.49.
994 * The fs.protected_regular and fs.protected_fifos sysctls, which were
995 added in Linux 4.19 to make some data spoofing attacks harder, are
996 now enabled by default. While this will hopefully improve the
997 security of most installations, it is technically a backwards
998 incompatible change; to disable these sysctls again, place the
999 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
1001 fs.protected_regular = 0
1002 fs.protected_fifos = 0
1004 Note that the similar hardlink and symlink protection has been
1005 enabled since v199, and may be disabled likewise.
1007 * The files read from the EnvironmentFile= setting in unit files now
1008 parse backslashes inside quotes literally, matching the behaviour of
1011 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
1012 now automatically become NOPs when run in a chroot() environment.
1014 * The tmpfiles.d/ "C" line type will now copy directory trees not only
1015 when the destination is so far missing, but also if it already exists
1016 as a directory and is empty. This is useful to cater for systems
1017 where directory trees are put together from multiple separate mount
1018 points but otherwise empty.
1020 * A new function sd_bus_close_unref() (and the associated
1021 sd_bus_close_unrefp()) has been added to libsystemd, that combines
1022 sd_bus_close() and sd_bus_unref() in one.
1024 * udevadm control learnt a new option for --ping for testing whether a
1025 systemd-udevd instance is running and reacting.
1027 * udevadm trigger learnt a new option for --wait-daemon for waiting
1028 systemd-udevd daemon to be initialized.
1030 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
1031 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
1032 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
1033 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
1034 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
1035 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
1036 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
1037 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
1038 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
1039 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
1040 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
1041 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
1042 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
1043 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
1044 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
1045 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
1046 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
1048 — Berlin, 2019-02-14
1052 * NoNewPrivileges=yes has been set for all long-running services
1053 implemented by systemd. Previously, this was problematic due to
1054 SELinux (as this would also prohibit the transition from PID1's label
1055 to the service's label). This restriction has since been lifted, but
1056 an SELinux policy update is required.
1057 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
1059 * DynamicUser=yes is dropped from systemd-networkd.service,
1060 systemd-resolved.service and systemd-timesyncd.service, which was
1061 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
1062 and since v236 for systemd-timesyncd.service. The users and groups
1063 systemd-network, systemd-resolve and systemd-timesync are created
1064 by systemd-sysusers again. Distributors or system administrators
1065 may need to create these users and groups if they not exist (or need
1066 to re-enable DynamicUser= for those units) while upgrading systemd.
1067 Also, the clock file for systemd-timesyncd may need to move from
1068 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
1070 * When unit files are loaded from disk, previously systemd would
1071 sometimes (depending on the unit loading order) load units from the
1072 target path of symlinks in .wants/ or .requires/ directories of other
1073 units. This meant that unit could be loaded from different paths
1074 depending on whether the unit was requested explicitly or as a
1075 dependency of another unit, not honouring the priority of directories
1076 in search path. It also meant that it was possible to successfully
1077 load and start units which are not found in the unit search path, as
1078 long as they were requested as a dependency and linked to from
1079 .wants/ or .requires/. The target paths of those symlinks are not
1080 used for loading units anymore and the unit file must be found in
1083 * A new service type has been added: Type=exec. It's very similar to
1084 Type=simple but ensures the service manager will wait for both fork()
1085 and execve() of the main service binary to complete before proceeding
1086 with follow-up units. This is primarily useful so that the manager
1087 propagates any errors in the preparation phase of service execution
1088 back to the job that requested the unit to be started. For example,
1089 consider a service that has ExecStart= set to a file system binary
1090 that doesn't exist. With Type=simple starting the unit would be
1091 considered instantly successful, as only fork() has to complete
1092 successfully and the manager does not wait for execve(), and hence
1093 its failure is seen "too late". With the new Type=exec service type
1094 starting the unit will fail, as the manager will wait for the
1095 execve() and notice its failure, which is then propagated back to the
1098 NOTE: with the next release 241 of systemd we intend to change the
1099 systemd-run tool to default to Type=exec for transient services
1100 started by it. This should be mostly safe, but in specific corner
1101 cases might result in problems, as the systemd-run tool will then
1102 block on NSS calls (such as user name look-ups due to User=) done
1103 between the fork() and execve(), which under specific circumstances
1104 might cause problems. It is recommended to specify "-p Type=simple"
1105 explicitly in the few cases where this applies. For regular,
1106 non-transient services (i.e. those defined with unit files on disk)
1107 we will continue to default to Type=simple.
1109 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
1110 userspace processes is set to 1024 (soft) and 4096
1111 (hard). Previously, systemd passed this on unmodified to all
1112 processes it forked off. With this systemd release the hard limit
1113 systemd passes on is increased to 512K, overriding the kernel's
1114 defaults and substantially increasing the number of simultaneous file
1115 descriptors unprivileged userspace processes can allocate. Note that
1116 the soft limit remains at 1024 for compatibility reasons: the
1117 traditional UNIX select() call cannot deal with file descriptors >=
1118 1024 and increasing the soft limit globally might thus result in
1119 programs unexpectedly allocating a high file descriptor and thus
1120 failing abnormally when attempting to use it with select() (of
1121 course, programs shouldn't use select() anymore, and prefer
1122 poll()/epoll, but the call unfortunately remains undeservedly popular
1123 at this time). This change reflects the fact that file descriptor
1124 handling in the Linux kernel has been optimized in more recent
1125 kernels and allocating large numbers of them should be much cheaper
1126 both in memory and in performance than it used to be. Programs that
1127 want to take benefit of the increased limit have to "opt-in" into
1128 high file descriptors explicitly by raising their soft limit. Of
1129 course, when they do that they must acknowledge that they cannot use
1130 select() anymore (and neither can any shared library they use — or
1131 any shared library used by any shared library they use and so on).
1132 Which default hard limit is most appropriate is of course hard to
1133 decide. However, given reports that ~300K file descriptors are used
1134 in real-life applications we believe 512K is sufficiently high as new
1135 default for now. Note that there are also reports that using very
1136 high hard limits (e.g. 1G) is problematic: some software allocates
1137 large arrays with one element for each potential file descriptor
1138 (Java, …) — a high hard limit thus triggers excessively large memory
1139 allocations in these applications. Hopefully, the new default of 512K
1140 is a good middle ground: higher than what real-life applications
1141 currently need, and low enough for avoid triggering excessively large
1142 allocations in problematic software. (And yes, somebody should fix
1145 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
1146 to the highest possible values, as separate accounting of file
1147 descriptors is no longer necessary, as memcg tracks them correctly as
1148 part of the memory accounting anyway. Thus, from the four limits on
1149 file descriptors currently enforced (fs.file-max, fs.nr_open,
1150 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
1151 and keep only the latter two. A set of build-time options
1152 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
1153 has been added to revert this change in behaviour, which might be
1154 an option for systems that turn off memcg in the kernel.
1156 * When no /etc/locale.conf file exists (and hence no locale settings
1157 are in place), systemd will now use the "C.UTF-8" locale by default,
1158 and set LANG= to it. This locale is supported by various
1159 distributions including Fedora, with clear indications that upstream
1160 glibc is going to make it available too. This locale enables UTF-8
1161 mode by default, which appears appropriate for 2018.
1163 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
1164 default. This effectively switches the RFC3704 Reverse Path filtering
1165 from Strict mode to Loose mode. This is more appropriate for hosts
1166 that have multiple links with routes to the same networks (e.g.
1167 a client with a Wi-Fi and Ethernet both connected to the internet).
1169 Consult the kernel documentation for details on this sysctl:
1170 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
1172 * CPUAccounting=yes no longer enables the CPU controller when using
1173 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
1174 statistics are now provided independently from the CPU controller.
1176 * Support for disabling a particular cgroup controller within a sub-tree
1177 has been added through the DisableControllers= directive.
1179 * cgroup_no_v1=all on the kernel command line now also implies
1180 using the unified cgroup hierarchy, unless one explicitly passes
1181 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
1183 * The new "MemoryMin=" unit file property may now be used to set the
1184 memory usage protection limit of processes invoked by the unit. This
1185 controls the cgroup v2 memory.min attribute. Similarly, the new
1186 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
1187 cgroup v2 io.latency cgroup property for configuring per-service I/O
1190 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
1191 to the cgroup v1 "devices" cgroup controller.
1193 * systemd-escape now is able to combine --unescape with --template. It
1194 also learnt a new option --instance for extracting and unescaping the
1195 instance part of a unit name.
1197 * sd-bus now provides the sd_bus_message_readv() which is similar to
1198 sd_bus_message_read() but takes a va_list object. The pair
1199 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
1200 has been added for configuring the default method call timeout to
1201 use. sd_bus_error_move() may be used to efficiently move the contents
1202 from one sd_bus_error structure to another, invalidating the
1203 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
1204 be used to control whether a bus connection object is automatically
1205 flushed when an sd-event loop is exited.
1207 * When processing classic BSD syslog log messages, journald will now
1208 save the original time-stamp string supplied in the new
1209 SYSLOG_TIMESTAMP= journal field. This permits consumers to
1210 reconstruct the original BSD syslog message more correctly.
1212 * StandardOutput=/StandardError= in service files gained support for
1213 new "append:…" parameters, for connecting STDOUT/STDERR of a service
1214 to a file, and appending to it.
1216 * The signal to use as last step of killing of unit processes is now
1217 configurable. Previously it was hard-coded to SIGKILL, which may now
1218 be overridden with the new KillSignal= setting. Note that this is the
1219 signal used when regular termination (i.e. SIGTERM) does not suffice.
1220 Similarly, the signal used when aborting a program in case of a
1221 watchdog timeout may now be configured too (WatchdogSignal=).
1223 * The XDG_SESSION_DESKTOP environment variable may now be configured in
1224 the pam_systemd argument line, using the new desktop= switch. This is
1225 useful to initialize it properly from a display manager without
1226 having to touch C code.
1228 * Most configuration options that previously accepted percentage values
1229 now also accept permille values with the '‰' suffix (instead of '%').
1231 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
1234 * systemd-resolved's configuration file resolved.conf gained a new
1235 option ReadEtcHosts= which may be used to turn off processing and
1236 honoring /etc/hosts entries.
1238 * The "--wait" switch may now be passed to "systemctl
1239 is-system-running", in which case the tool will synchronously wait
1240 until the system finished start-up.
1242 * hostnamed gained a new bus call to determine the DMI product UUID.
1244 * On x86-64 systemd will now prefer using the RDRAND processor
1245 instruction over /dev/urandom whenever it requires randomness that
1246 neither has to be crypto-grade nor should be reproducible. This
1247 should substantially reduce the amount of entropy systemd requests
1248 from the kernel during initialization on such systems, though not
1249 reduce it to zero. (Why not zero? systemd still needs to allocate
1250 UUIDs and such uniquely, which require high-quality randomness.)
1252 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
1253 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
1254 for forcing the "Other Information" bit in IPv6 RA messages. The
1255 bonding logic gained four new options AdActorSystemPriority=,
1256 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
1257 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
1258 shuffling of flows. The tunnel logic gained a new
1259 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
1260 Deployment. The policy rule logic gained four new options IPProtocol=,
1261 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
1262 support for the MulticastToUnicast= option. networkd also gained
1263 support for configuring static IPv4 ARP or IPv6 neighbor entries.
1265 * .preset files (as read by 'systemctl preset') may now be used to
1266 instantiate services.
1268 * /etc/crypttab now understands the sector-size= option to configure
1269 the sector size for an encrypted partition.
1271 * Key material for encrypted disks may now be placed on a formatted
1272 medium, and referenced from /etc/crypttab by the UUID of the file
1273 system, followed by "=" suffixed by the path to the key file.
1275 * The "collect" udev component has been removed without replacement, as
1276 it is neither used nor maintained.
1278 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
1279 LogsDirectory=, ConfigurationDirectory= settings are used in a
1280 service the executed processes will now receive a set of environment
1281 variables containing the full paths of these directories.
1282 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
1283 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
1284 are used. Note that these options may be used multiple times per
1285 service in which case the resulting paths will be concatenated and
1286 separated by colons.
1288 * Predictable interface naming has been extended to cover InfiniBand
1289 NICs. They will be exposed with an "ib" prefix.
1291 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
1292 which case the respective line failing is ignored.
1294 * .link files may now be used to configure the equivalent to the
1295 "ethtool advertise" commands.
1297 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
1298 alternative to libudev.h. Previously, the latter was just an internal
1299 wrapper around the former, but now these two APIs are exposed
1302 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
1303 which calculates an app-specific boot ID similar to how
1304 sd_id128_get_machine_app_specific() generates an app-specific machine
1307 * A new tool systemd-id128 has been added that can be used to determine
1308 and generate various 128bit IDs.
1310 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
1313 * systemd-hibernate-resume-generator will now honor the "noresume"
1314 kernel command line option, in which case it will bypass resuming
1315 from any hibernated image.
1317 * The systemd-sleep.conf configuration file gained new options
1318 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
1319 AllowHybridSleep= for prohibiting specific sleep modes even if the
1320 kernel exports them.
1322 * portablectl is now officially supported and has thus moved to
1325 * bootctl learnt the two new commands "set-default" and "set-oneshot"
1326 for setting the default boot loader item to boot to (either
1327 persistently or only for the next boot). This is currently only
1328 compatible with sd-boot, but may be implemented on other boot loaders
1329 too, that follow the boot loader interface. The updated interface is
1330 now documented here:
1332 https://systemd.io/BOOT_LOADER_INTERFACE
1334 * A new kernel command line option systemd.early_core_pattern= is now
1335 understood which may be used to influence the core_pattern PID 1
1336 installs during early boot.
1338 * busctl learnt two new options -j and --json= for outputting method
1339 call replies, properties and monitoring output in JSON.
1341 * journalctl's JSON output now supports simple ANSI coloring as well as
1342 a new "json-seq" mode for generating RFC7464 output.
1344 * Unit files now support the %g/%G specifiers that resolve to the UNIX
1345 group/GID of the service manager runs as, similar to the existing
1346 %u/%U specifiers that resolve to the UNIX user/UID.
1348 * systemd-logind learnt a new global configuration option
1349 UserStopDelaySec= that may be set in logind.conf. It specifies how
1350 long the systemd --user instance shall remain started after a user
1351 logs out. This is useful to speed up repetitive re-connections of the
1352 same user, as it means the user's service manager doesn't have to be
1353 stopped/restarted on each iteration, but can be reused between
1354 subsequent options. This setting defaults to 10s. systemd-logind also
1355 exports two new properties on its Manager D-Bus objects indicating
1356 whether the system's lid is currently closed, and whether the system
1359 * systemd gained support for a generic boot counting logic, which
1360 generically permits automatic reverting to older boot loader entries
1361 if newer updated ones don't work. The boot loader side is implemented
1362 in sd-boot, but is kept open for other boot loaders too. For details
1365 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
1367 * The SuccessAction=/FailureAction= unit file settings now learnt two
1368 new parameters: "exit" and "exit-force", which result in immediate
1369 exiting of the service manager, and are only useful in systemd --user
1370 and container environments.
1372 * Unit files gained support for a pair of options
1373 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
1374 exit status to use as service manager exit status when
1375 SuccessAction=/FailureAction= is set to exit or exit-force.
1377 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
1378 options may now be used to configure the log rate limiting applied by
1379 journald per-service.
1381 * systemd-analyze gained a new verb "timespan" for parsing and
1382 normalizing time span values (i.e. strings like "5min 7s 8us").
1384 * systemd-analyze also gained a new verb "security" for analyzing the
1385 security and sand-boxing settings of services in order to determine an
1386 "exposure level" for them, indicating whether a service would benefit
1387 from more sand-boxing options turned on for them.
1389 * "systemd-analyze syscall-filter" will now also show system calls
1390 supported by the local kernel but not included in any of the defined
1393 * .nspawn files now understand the Ephemeral= setting, matching the
1394 --ephemeral command line switch.
1396 * sd-event gained the new APIs sd_event_source_get_floating() and
1397 sd_event_source_set_floating() for controlling whether a specific
1398 event source is "floating", i.e. destroyed along with the even loop
1401 * Unit objects on D-Bus gained a new "Refs" property that lists all
1402 clients that currently have a reference on the unit (to ensure it is
1405 * The JoinControllers= option in system.conf is no longer supported, as
1406 it didn't work correctly, is hard to support properly, is legacy (as
1407 the concept only exists on cgroup v1) and apparently wasn't used.
1409 * Journal messages that are generated whenever a unit enters the failed
1410 state are now tagged with a unique MESSAGE_ID. Similarly, messages
1411 generated whenever a service process exits are now made recognizable,
1412 too. A tagged message is also emitted whenever a unit enters the
1413 "dead" state on success.
1415 * systemd-run gained a new switch --working-directory= for configuring
1416 the working directory of the service to start. A shortcut -d is
1417 equivalent, setting the working directory of the service to the
1418 current working directory of the invoking program. The new --shell
1419 (or just -S) option has been added for invoking the $SHELL of the
1420 caller as a service, and implies --pty --same-dir --wait --collect
1421 --service-type=exec. Or in other words, "systemd-run -S" is now the
1422 quickest way to quickly get an interactive in a fully clean and
1423 well-defined system service context.
1425 * machinectl gained a new verb "import-fs" for importing an OS tree
1426 from a directory. Moreover, when a directory or tarball is imported
1427 and single top-level directory found with the OS itself below the OS
1428 tree is automatically mangled and moved one level up.
1430 * systemd-importd will no longer set up an implicit btrfs loop-back
1431 file system on /var/lib/machines. If one is already set up, it will
1432 continue to be used.
1434 * A new generator "systemd-run-generator" has been added. It will
1435 synthesize a unit from one or more program command lines included in
1436 the kernel command line. This is very useful in container managers
1439 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
1441 This will run "systemd-nspawn" on an image, invoke the specified
1442 command line and immediately shut down the container again, returning
1443 the command line's exit code.
1445 * The block device locking logic is now documented:
1447 https://systemd.io/BLOCK_DEVICE_LOCKING
1449 * loginctl and machinectl now optionally output the various tables in
1450 JSON using the --output= switch. It is our intention to add similar
1451 support to systemctl and all other commands.
1453 * udevadm's query and trigger verb now optionally take a .device unit
1456 * systemd-udevd's network naming logic now understands a new
1457 net.naming-scheme= kernel command line switch, which may be used to
1458 pick a specific version of the naming scheme. This helps stabilizing
1459 interface names even as systemd/udev are updated and the naming logic
1462 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
1463 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
1464 initialize one to all 0xFF.
1466 * After loading the SELinux policy systemd will now recursively relabel
1467 all files and directories listed in
1468 /run/systemd/relabel-extra.d/*.relabel (which should be simple
1469 newline separated lists of paths) in addition to the ones it already
1470 implicitly relabels in /run, /dev and /sys. After the relabelling is
1471 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
1472 removed. This is useful to permit initrds (i.e. code running before
1473 the SELinux policy is in effect) to generate files in the host
1474 filesystem safely and ensure that the correct label is applied during
1475 the transition to the host OS.
1477 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
1478 mknod() handling in user namespaces. Previously mknod() would always
1479 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
1480 but device nodes generated that way cannot be opened, and attempts to
1481 open them result in EPERM. This breaks the "graceful fallback" logic
1482 in systemd's PrivateDevices= sand-boxing option. This option is
1483 implemented defensively, so that when systemd detects it runs in a
1484 restricted environment (such as a user namespace, or an environment
1485 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
1486 where device nodes cannot be created the effect of PrivateDevices= is
1487 bypassed (following the logic that 2nd-level sand-boxing is not
1488 essential if the system systemd runs in is itself already sand-boxed
1489 as a whole). This logic breaks with 4.18 in container managers where
1490 user namespacing is used: suddenly PrivateDevices= succeeds setting
1491 up a private /dev/ file system containing devices nodes — but when
1492 these are opened they don't work.
1494 At this point it is recommended that container managers utilizing
1495 user namespaces that intend to run systemd in the payload explicitly
1496 block mknod() with seccomp or similar, so that the graceful fallback
1499 We are very sorry for the breakage and the requirement to change
1500 container configurations for newer kernels. It's purely caused by an
1501 incompatible kernel change. The relevant kernel developers have been
1502 notified about this userspace breakage quickly, but they chose to
1505 * PermissionsStartOnly= setting is deprecated (but is still supported
1506 for backwards compatibility). The same functionality is provided by
1507 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
1510 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
1511 pam_systemd anymore.
1513 * The naming scheme for network devices was changed to always rename
1514 devices, even if they were already renamed by userspace. The "kernel"
1515 policy was changed to only apply as a fallback, if no other naming
1518 * The requirements to build systemd is bumped to meson-0.46 and
1521 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
1522 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
1523 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
1524 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
1525 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
1526 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
1527 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
1528 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
1529 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
1530 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
1531 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
1532 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
1533 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
1534 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
1535 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
1536 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
1537 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
1538 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
1539 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
1540 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
1541 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
1542 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
1543 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
1544 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
1545 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
1546 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
1547 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
1548 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
1549 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
1550 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
1551 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
1552 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
1553 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
1554 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
1555 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
1556 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
1557 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
1558 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
1559 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
1560 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
1561 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
1562 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
1563 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
1564 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
1565 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
1567 — Warsaw, 2018-12-21
1571 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
1572 builtin will name network interfaces differently than in previous
1573 versions for virtual network interfaces created with SR-IOV and NPAR
1574 and for devices where the PCI network controller device does not have
1575 a slot number associated.
1577 SR-IOV virtual devices are now named based on the name of the parent
1578 interface, with a suffix of "v<N>", where <N> is the virtual device
1579 number. Previously those virtual devices were named as if completely
1582 The ninth and later NPAR virtual devices will be named following the
1583 scheme used for the first eight NPAR partitions. Previously those
1584 devices were not renamed and the kernel default (eth<n>) was used.
1586 "net_id" will also generate names for PCI devices where the PCI
1587 network controller device does not have an associated slot number
1588 itself, but one of its parents does. Previously those devices were
1589 not renamed and the kernel default (eth<n>) was used.
1591 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
1592 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
1593 the unit. So, it is expected that the default behavior of
1594 systemd-logind is not changed. However, if distribution packagers or
1595 administrators disabled or modified IPAddressDeny= setting by a
1596 drop-in config file, then it may be necessary to update the file to
1597 re-enable AF_INET and AF_INET6 to support network user name services,
1600 * When the RestrictNamespaces= unit property is specified multiple
1601 times, then the specified types are merged now. Previously, only the
1602 last assignment was used. So, if distribution packagers or
1603 administrators modified the setting by a drop-in config file, then it
1604 may be necessary to update the file.
1606 * When OnFailure= is used in combination with Restart= on a service
1607 unit, then the specified units will no longer be triggered on
1608 failures that result in restarting. Previously, the specified units
1609 would be activated each time the unit failed, even when the unit was
1610 going to be restarted automatically. This behaviour contradicted the
1611 documentation. With this release the code is adjusted to match the
1614 * systemd-tmpfiles will now print a notice whenever it encounters
1615 tmpfiles.d/ lines referencing the /var/run/ directory. It will
1616 recommend reworking them to use the /run/ directory instead (for
1617 which /var/run/ is simply a symlinked compatibility alias). This way
1618 systemd-tmpfiles can properly detect line conflicts and merge lines
1619 referencing the same file by two paths, without having to access
1622 * systemctl disable/unmask/preset/preset-all cannot be used with
1623 --runtime. Previously this was allowed, but resulted in unintuitive
1624 behaviour that wasn't useful. systemctl disable/unmask will now undo
1625 both runtime and persistent enablement/masking, i.e. it will remove
1626 any relevant symlinks both in /run and /etc.
1628 * Note that all long-running system services shipped with systemd will
1629 now default to a system call whitelist (rather than a blacklist, as
1630 before). In particular, systemd-udevd will now enforce one too. For
1631 most cases this should be safe, however downstream distributions
1632 which disabled sandboxing of systemd-udevd (specifically the
1633 MountFlags= setting), might want to disable this security feature
1634 too, as the default whitelisting will prohibit all mount, swap,
1635 reboot and clock changing operations from udev rules.
1637 * sd-boot acquired new loader configuration settings to optionally turn
1638 off Windows and MacOS boot partition discovery as well as
1639 reboot-into-firmware menu items. It is also able to pick a better
1640 screen resolution for HiDPI systems, and now provides loader
1641 configuration settings to change the resolution explicitly.
1643 * systemd-resolved now supports DNS-over-TLS. It's still
1644 turned off by default, use DNSOverTLS=opportunistic to turn it on in
1645 resolved.conf. We intend to make this the default as soon as couple
1646 of additional techniques for optimizing the initial latency caused by
1647 establishing a TLS/TCP connection are implemented.
1649 * systemd-resolved.service and systemd-networkd.service now set
1650 DynamicUser=yes. The users systemd-resolve and systemd-network are
1651 not created by systemd-sysusers anymore.
1653 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
1654 that embed a network facing module into any process using getpwuid()
1655 or related call: the dynamic allocation of the user ID for
1656 systemd-resolved.service means the service manager has to check NSS
1657 if the user name is already taken when forking off the service. Since
1658 the user in the common case won't be defined in /etc/passwd the
1659 lookup is likely to trigger nss-ldap which in turn might use NSS to
1660 ask systemd-resolved for hostname lookups. This will hence result in
1661 a deadlock: a user name lookup in order to start
1662 systemd-resolved.service will result in a host name lookup for which
1663 systemd-resolved.service needs to be started already. There are
1664 multiple ways to work around this problem: pre-allocate the
1665 "systemd-resolve" user on such systems, so that nss-ldap won't be
1666 triggered; or use a different NSS package that doesn't do networking
1667 in-process but provides a local asynchronous name cache; or configure
1668 the NSS package to avoid lookups for UIDs in the range `pkg-config
1669 systemd --variable=dynamicuidmin` … `pkg-config systemd
1670 --variable=dynamicuidmax`, so that it does not consider itself
1671 authoritative for the same UID range systemd allocates dynamic users
1674 * The systemd-resolve tool has been renamed to resolvectl (it also
1675 remains available under the old name, for compatibility), and its
1676 interface is now verb-based, similar in style to the other <xyz>ctl
1677 tools, such as systemctl or loginctl.
1679 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
1680 compatibility. It may be symlinked under the 'resolvconf' name, in
1681 which case it will take arguments and input compatible with the
1682 Debian and FreeBSD resolvconf tool.
1684 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
1685 where the system initially suspends, and after a timeout resumes and
1688 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
1689 set the client will only send a DUID as client identifier.
1691 * The nss-systemd glibc NSS module will now enumerate dynamic users and
1692 groups in effect. Previously, it could resolve UIDs/GIDs to user
1693 names/groups and vice versa, but did not support enumeration.
1695 * journald's Compress= configuration setting now optionally accepts a
1696 byte threshold value. All journal objects larger than this threshold
1697 will be compressed, smaller ones will not. Previously this threshold
1698 was not configurable and set to 512.
1700 * A new system.conf setting NoNewPrivileges= is now available which may
1701 be used to turn off acquisition of new privileges system-wide
1702 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
1703 for all its children). Note that turning this option on means setuid
1704 binaries and file system capabilities lose their special powers.
1705 While turning on this option is a big step towards a more secure
1706 system, doing so is likely to break numerous pre-existing UNIX tools,
1707 in particular su and sudo.
1709 * A new service systemd-time-sync-wait.service has been added. If
1710 enabled it will delay the time-sync.target unit at boot until time
1711 synchronization has been received from the network. This
1712 functionality is useful on systems lacking a local RTC or where it is
1713 acceptable that the boot process shall be delayed by external network
1716 * When hibernating, systemd will now inform the kernel of the image
1717 write offset, on kernels new enough to support this. This means swap
1718 files should work for hibernation now.
1720 * When loading unit files, systemd will now look for drop-in unit files
1721 extensions in additional places. Previously, for a unit file name
1722 "foo-bar-baz.service" it would look for dropin files in
1723 "foo-bar-baz.service.d/*.conf". Now, it will also look in
1724 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
1725 service name truncated after all inner dashes. This scheme allows
1726 writing drop-ins easily that apply to a whole set of unit files at
1727 once. It's particularly useful for mount and slice units (as their
1728 naming is prefix based), but is also useful for service and other
1729 units, for packages that install multiple unit files at once,
1730 following a strict naming regime of beginning the unit file name with
1731 the package's name. Two new specifiers are now supported in unit
1732 files to match this: %j and %J are replaced by the part of the unit
1733 name following the last dash.
1735 * Unit files and other configuration files that support specifier
1736 expansion now understand another three new specifiers: %T and %V will
1737 resolve to /tmp and /var/tmp respectively, or whatever temporary
1738 directory has been set for the calling user. %E will expand to either
1739 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
1741 * The ExecStart= lines of unit files are no longer required to
1742 reference absolute paths. If non-absolute paths are specified the
1743 specified binary name is searched within the service manager's
1744 built-in $PATH, which may be queried with 'systemd-path
1745 search-binaries-default'. It's generally recommended to continue to
1746 use absolute paths for all binaries specified in unit files.
1748 * Units gained a new load state "bad-setting", which is used when a
1749 unit file was loaded, but contained fatal errors which prevent it
1750 from being started (for example, a service unit has been defined
1751 lacking both ExecStart= and ExecStop= lines).
1753 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
1754 support alternative debuggers, for example lldb. The old name
1755 continues to be available however, for compatibility reasons. Use the
1756 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
1757 to pick an alternative debugger instead of the default gdb.
1759 * systemctl and the other tools will now output escape sequences that
1760 generate proper clickable hyperlinks in various terminal emulators
1761 where useful (for example, in the "systemctl status" output you can
1762 now click on the unit file name to quickly open it in the
1763 editor/viewer of your choice). Note that not all terminal emulators
1764 support this functionality yet, but many do. Unfortunately, the
1765 "less" pager doesn't support this yet, hence this functionality is
1766 currently automatically turned off when a pager is started (which
1767 happens quite often due to auto-paging). We hope to remove this
1768 limitation as soon as "less" learns these escape sequences. This new
1769 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
1770 environment variable. For details on these escape sequences see:
1771 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
1773 * networkd's .network files now support a new IPv6MTUBytes= option for
1774 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
1775 option in the [Route] section to configure the MTU to use for
1776 specific routes. It also gained support for configuration of the DHCP
1777 "UserClass" option through the new UserClass= setting. It gained
1778 three new options in the new [CAN] section for configuring CAN
1779 networks. The MULTICAST and ALLMULTI interface flags may now be
1780 controlled explicitly with the new Multicast= and AllMulticast=
1783 * networkd will now automatically make use of the kernel's route
1784 expiration feature, if it is available.
1786 * udevd's .link files now support setting the number of receive and
1787 transmit channels, using the RxChannels=, TxChannels=,
1788 OtherChannels=, CombinedChannels= settings.
1790 * Support for UDPSegmentationOffload= has been removed, given its
1791 limited support in hardware, and waning software support.
1793 * networkd's .netdev files now support creating "netdevsim" interfaces.
1795 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
1796 to query the unit belonging to a specific kernel control group.
1798 * systemd-analyze gained a new verb "cat-config", which may be used to
1799 dump the contents of any configuration file, with all its matching
1800 drop-in files added in, and honouring the usual search and masking
1801 logic applied to systemd configuration files. For example use
1802 "systemd-analyze cat-config systemd/system.conf" to get the complete
1803 system configuration file of systemd how it would be loaded by PID 1
1804 itself. Similar to this, various tools such as systemd-tmpfiles or
1805 systemd-sysusers, gained a new option "--cat-config", which does the
1806 corresponding operation for their own configuration settings. For
1807 example, "systemd-tmpfiles --cat-config" will now output the full
1808 list of tmpfiles.d/ lines in place.
1810 * timedatectl gained three new verbs: "show" shows bus properties of
1811 systemd-timedated, "timesync-status" shows the current NTP
1812 synchronization state of systemd-timesyncd, and "show-timesync"
1813 shows bus properties of systemd-timesyncd.
1815 * systemd-timesyncd gained a bus interface on which it exposes details
1818 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
1819 understood by systemd-timedated. It takes a colon-separated list of
1820 unit names of NTP client services. The list is used by
1821 "timedatectl set-ntp".
1823 * systemd-nspawn gained a new --rlimit= switch for setting initial
1824 resource limits for the container payload. There's a new switch
1825 --hostname= to explicitly override the container's hostname. A new
1826 --no-new-privileges= switch may be used to control the
1827 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
1828 --oom-score-adjust= switch controls the OOM scoring adjustment value
1829 for the payload. The new --cpu-affinity= switch controls the CPU
1830 affinity of the container payload. The new --resolv-conf= switch
1831 allows more detailed control of /etc/resolv.conf handling of the
1832 container. Similarly, the new --timezone= switch allows more detailed
1833 control of /etc/localtime handling of the container.
1835 * systemd-detect-virt gained a new --list switch, which will print a
1836 list of all currently known VM and container environments.
1838 * Support for "Portable Services" has been added, see
1839 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
1840 experimental, but this is expected to change soon. Reflecting this
1841 experimental state, the "portablectl" binary is not installed into
1842 /usr/bin yet. The binary has to be called with the full path
1843 /usr/lib/systemd/portablectl instead.
1845 * journalctl's and systemctl's -o switch now knows a new log output
1846 mode "with-unit". The output it generates is very similar to the
1847 regular "short" mode, but displays the unit name instead of the
1848 syslog tag for each log line. Also, the date is shown with timezone
1849 information. This mode is probably more useful than the classic
1850 "short" output mode for most purposes, except where pixel-perfect
1851 compatibility with classic /var/log/messages formatting is required.
1853 * A new --dump-bus-properties switch has been added to the systemd
1854 binary, which may be used to dump all supported D-Bus properties.
1855 (Options which are still supported, but are deprecated, are *not*
1858 * sd-bus gained a set of new calls:
1859 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
1860 enable/disable the "floating" state of a bus slot object,
1861 i.e. whether the slot object pins the bus it is allocated for into
1862 memory or if the bus slot object gets disconnected when the bus goes
1863 away. sd_bus_open_with_description(),
1864 sd_bus_open_user_with_description(),
1865 sd_bus_open_system_with_description() may be used to allocate bus
1866 objects and set their description string already during allocation.
1868 * sd-event gained support for watching inotify events from the event
1869 loop, in an efficient way, sharing inotify handles between multiple
1870 users. For this a new function sd_event_add_inotify() has been added.
1872 * sd-event and sd-bus gained support for calling special user-supplied
1873 destructor functions for userdata pointers associated with
1874 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
1875 functions sd_bus_slot_set_destroy_callback,
1876 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
1877 sd_bus_track_get_destroy_callback,
1878 sd_event_source_set_destroy_callback,
1879 sd_event_source_get_destroy_callback have been added.
1881 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
1883 * PID 1 will now automatically reschedule .timer units whenever the
1884 local timezone changes. (They previously got rescheduled
1885 automatically when the system clock changed.)
1887 * New documentation has been added to document cgroups delegation,
1888 portable services and the various code quality tools we have set up:
1890 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
1891 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
1892 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
1894 * The Boot Loader Specification has been added to the source tree.
1896 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
1898 While moving it into our source tree we have updated it and further
1899 changes are now accepted through the usual github PR workflow.
1901 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
1902 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
1903 earlier PAM modules. The data in these fields is used to initialize
1904 the session scope's resource properties. Thus external PAM modules
1905 may now configure per-session limits, for example sourced from
1906 external user databases.
1908 * socket units with Accept=yes will now maintain a "refused" counter in
1909 addition to the existing "accepted" counter, counting connections
1910 refused due to the enforced limits.
1912 * The "systemd-path search-binaries-default" command may now be use to
1913 query the default, built-in $PATH PID 1 will pass to the services it
1916 * A new unit file setting PrivateMounts= has been added. It's a boolean
1917 option. If enabled the unit's processes are invoked in their own file
1918 system namespace. Note that this behaviour is also implied if any
1919 other file system namespacing options (such as PrivateTmp=,
1920 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
1921 primarily useful for services that do not use any of the other file
1922 system namespacing options. One such service is systemd-udevd.service
1923 where this is now used by default.
1925 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
1926 when the system is booted in UEFI "secure mode".
1928 * A new unit "system-update-pre.target" is added, which defines an
1929 optional synchronization point for offline system updates, as
1930 implemented by the pre-existing "system-update.target" unit. It
1931 allows ordering services before the service that executes the actual
1932 update process in a generic way.
1934 * Systemd now emits warnings whenever .include syntax is used.
1936 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
1937 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
1938 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
1939 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
1940 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
1941 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
1942 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
1943 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
1944 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
1945 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
1946 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
1947 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
1948 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
1949 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
1950 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
1951 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
1952 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
1953 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
1954 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
1955 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
1956 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
1957 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
1958 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
1959 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
1960 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
1961 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
1962 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
1963 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
1964 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1966 — Berlin, 2018-06-22
1970 * The MemoryAccounting= unit property now defaults to on. After
1971 discussions with the upstream control group maintainers we learnt
1972 that the negative impact of cgroup memory accounting on current
1973 kernels is finally relatively minimal, so that it should be safe to
1974 enable this by default without affecting system performance. Besides
1975 memory accounting only task accounting is turned on by default, all
1976 other forms of resource accounting (CPU, IO, IP) remain off for now,
1977 because it's not clear yet that their impact is small enough to move
1978 from opt-in to opt-out. We recommend downstreams to leave memory
1979 accounting on by default if kernel 4.14 or higher is primarily
1980 used. On very resource constrained systems or when support for old
1981 kernels is a necessity, -Dmemory-accounting-default=false can be used
1982 to revert this change.
1984 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
1985 %udev_rules_update) and the journal catalog (%journal_catalog_update)
1986 from the upgrade scriptlets of individual packages now do nothing.
1987 Transfiletriggers have been added which will perform those updates
1988 once at the end of the transaction.
1990 Similar transfiletriggers have been added to execute any sysctl.d
1991 and binfmt.d rules. Thus, it should be unnecessary to provide any
1992 scriptlets to execute this configuration from package installation
1995 * systemd-sysusers gained a mode where the configuration to execute is
1996 specified on the command line, but this configuration is not executed
1997 directly, but instead it is merged with the configuration on disk,
1998 and the result is executed. This is useful for package installation
1999 scripts which want to create the user before installing any files on
2000 disk (in case some of those files are owned by that user), while
2001 still allowing local admin overrides.
2003 This functionality is exposed to rpm scriptlets through a new
2004 %sysusers_create_package macro. Old %sysusers_create and
2005 %sysusers_create_inline macros are deprecated.
2007 A transfiletrigger for sysusers.d configuration is now installed,
2008 which means that it should be unnecessary to call systemd-sysusers from
2009 package installation scripts, unless the package installs any files
2010 owned by those newly-created users, in which case
2011 %sysusers_create_package should be used.
2013 * Analogous change has been done for systemd-tmpfiles: it gained a mode
2014 where the command-line configuration is merged with the configuration
2015 on disk. This is exposed as the new %tmpfiles_create_package macro,
2016 and %tmpfiles_create is deprecated. A transfiletrigger is installed
2017 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
2018 from package installation scripts.
2020 * sysusers.d configuration for a user may now also specify the group
2021 number, in addition to the user number ("u username 123:456"), or
2022 without the user number ("u username -:456").
2024 * Configution items for systemd-sysusers can now be specified as
2025 positional arguments when the new --inline switch is used.
2027 * The login shell of users created through sysusers.d may now be
2028 specified (previously, it was always /bin/sh for root and
2029 /sbin/nologin for other users).
2031 * systemd-analyze gained a new --global switch to look at global user
2032 configuration. It also gained a unit-paths verb to list the unit load
2033 paths that are compiled into systemd (which can be used with
2034 --systemd, --user, or --global).
2036 * udevadm trigger gained a new --settle/-w option to wait for any
2037 triggered events to finish (but just those, and not any other events
2038 which are triggered meanwhile).
2040 * The action that systemd-logind takes when the lid is closed and the
2041 machine is connected to external power can now be configured using
2042 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
2043 was determined by HandleLidSwitch=, and, for backwards compatibility,
2044 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
2046 * journalctl will periodically call sd_journal_process() to make it
2047 resilient against inotify queue overruns when journal files are
2048 rotated very quickly.
2050 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
2051 sd_bus_get_n_queued_write — may be used to check the number of
2052 pending bus messages.
2054 * systemd gained a new
2055 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
2056 which can be used to migrate foreign processes to scope and service
2057 units. The primary user for this new API is systemd itself: the
2058 systemd --user instance uses this call of the systemd --system
2059 instance to migrate processes if it itself gets the request to
2060 migrate processes and the kernel refuses this due to access
2061 restrictions. Thanks to this "systemd-run --scope --user …" works
2062 again in pure cgroup v2 environments when invoked from the user
2065 * A new TemporaryFileSystem= setting can be used to mask out part of
2066 the real file system tree with tmpfs mounts. This may be combined
2067 with BindPaths= and BindReadOnlyPaths= to hide files or directories
2068 not relevant to the unit, while still allowing some paths lower in
2069 the tree to be accessed.
2071 ProtectHome=tmpfs may now be used to hide user home and runtime
2072 directories from units, in a way that is mostly equivalent to
2073 "TemporaryFileSystem=/home /run/user /root".
2075 * Non-service units are now started with KeyringMode=shared by default.
2076 This means that mount and swapon and other mount tools have access
2077 to keys in the main keyring.
2079 * /sys/fs/bpf is now mounted automatically.
2081 * QNX virtualization is now detected by systemd-detect-virt and may
2082 be used in ConditionVirtualization=.
2084 * IPAccounting= may now be enabled also for slice units.
2086 * A new -Dsplit-bin= build configuration switch may be used to specify
2087 whether bin and sbin directories are merged, or if they should be
2088 included separately in $PATH and various listings of executable
2089 directories. The build configuration scripts will try to autodetect
2090 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
2091 system, but distributions are encouraged to configure this
2094 * A new -Dok-color= build configuration switch may be used to change
2095 the colour of "OK" status messages.
2097 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
2098 PrivateNetwork=yes was buggy in previous versions of systemd. This
2099 means that after the upgrade and daemon-reexec, any such units must
2102 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
2103 will not exclude read-only files owned by root from cleanup.
2105 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
2106 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
2107 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
2108 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
2109 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
2110 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
2111 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
2112 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
2113 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
2114 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
2115 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
2116 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
2117 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
2118 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
2119 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
2120 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
2122 — Warsaw, 2018-03-05
2126 * Some keyboards come with a zoom see-saw or rocker which until now got
2127 mapped to the Linux "zoomin/out" keys in hwdb. However, these
2128 keycodes are not recognized by any major desktop. They now produce
2129 Up/Down key events so that they can be used for scrolling.
2131 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
2132 slightly: previously, if an argument was specified for lines of this
2133 type (i.e. the right-most column was set) this string was appended to
2134 existing files each time systemd-tmpfiles was run. This behaviour was
2135 different from what the documentation said, and not particularly
2136 useful, as repeated systemd-tmpfiles invocations would not be
2137 idempotent and grow such files without bounds. With this release
2138 behaviour has been altered to match what the documentation says:
2139 lines of this type only have an effect if the indicated files don't
2140 exist yet, and only then the argument string is written to the file.
2142 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
2143 systemd-tmpfiles behaviour: previously, read-only files owned by root
2144 were always excluded from the file "aging" algorithm (i.e. the
2145 automatic clean-up of directories like /tmp based on
2146 atime/mtime/ctime). We intend to drop this restriction, and age files
2147 by default even when owned by root and read-only. This behaviour was
2148 inherited from older tools, but there have been requests to remove
2149 it, and it's not obvious why this restriction was made in the first
2150 place. Please speak up now, if you are aware of software that reqires
2151 this behaviour, otherwise we'll remove the restriction in v238.
2153 * A new environment variable $SYSTEMD_OFFLINE is now understood by
2154 systemctl. It takes a boolean argument. If on, systemctl assumes it
2155 operates on an "offline" OS tree, and will not attempt to talk to the
2156 service manager. Previously, this mode was implicitly enabled if a
2157 chroot() environment was detected, and this new environment variable
2158 now provides explicit control.
2160 * .path and .socket units may now be created transiently, too.
2161 Previously only service, mount, automount and timer units were
2162 supported as transient units. The systemd-run tool has been updated
2163 to expose this new functionality, you may hence use it now to bind
2164 arbitrary commands to path or socket activation on-the-fly from the
2165 command line. Moreover, almost all properties are now exposed for the
2166 unit types that already supported transient operation.
2168 * The systemd-mount command gained support for a new --owner= parameter
2169 which takes a user name, which is then resolved and included in uid=
2170 and gid= mount options string of the file system to mount.
2172 * A new unit condition ConditionControlGroupController= has been added
2173 that checks whether a specific cgroup controller is available.
2175 * Unit files, udev's .link files, and systemd-networkd's .netdev and
2176 .network files all gained support for a new condition
2177 ConditionKernelVersion= for checking against specific kernel
2180 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
2181 support for configuring device flags in the Flags= setting. In the
2182 same files, the [Tunnel] section gained support for configuring
2183 AllowLocalRemote=. The [Route] section in .network files gained
2184 support for configuring InitialCongestionWindow=,
2185 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
2186 understands RapidCommit=.
2188 * systemd-networkd's DHCPv6 support gained support for Prefix
2191 * sd-bus gained support for a new "watch-bind" feature. When this
2192 feature is enabled, an sd_bus connection may be set up to connect to
2193 an AF_UNIX socket in the file system as soon as it is created. This
2194 functionality is useful for writing early-boot services that
2195 automatically connect to the system bus as soon as it is started,
2196 without ugly time-based polling. systemd-networkd and
2197 systemd-resolved have been updated to make use of this
2198 functionality. busctl exposes this functionality in a new
2199 --watch-bind= command line switch.
2201 * sd-bus will now optionally synthesize a local "Connected" signal as
2202 soon as a D-Bus connection is set up fully. This message mirrors the
2203 already existing "Disconnected" signal which is synthesized when the
2204 connection is terminated. This signal is generally useful but
2205 particularly handy in combination with the "watch-bind" feature
2206 described above. Synthesizing of this message has to be requested
2207 explicitly through the new API call sd_bus_set_connected_signal(). In
2208 addition a new call sd_bus_is_ready() has been added that checks
2209 whether a connection is fully set up (i.e. between the "Connected" and
2210 "Disconnected" signals).
2212 * sd-bus gained two new calls sd_bus_request_name_async() and
2213 sd_bus_release_name_async() for asynchronously registering bus
2214 names. Similar, there is now sd_bus_add_match_async() for installing
2215 a signal match asynchronously. All of systemd's own services have
2216 been updated to make use of these calls. Doing these operations
2217 asynchronously has two benefits: it reduces the risk of deadlocks in
2218 case of cyclic dependencies between bus services, and it speeds up
2219 service initialization since synchronization points for bus
2220 round-trips are removed.
2222 * sd-bus gained two new calls sd_bus_match_signal() and
2223 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
2224 and sd_bus_add_match_async() but instead of taking a D-Bus match
2225 string take match fields as normal function parameters.
2227 * sd-bus gained two new calls sd_bus_set_sender() and
2228 sd_bus_message_set_sender() for setting the sender name of outgoing
2229 messages (either for all outgoing messages or for just one specific
2230 one). These calls are only useful in direct connections as on
2231 brokered connections the broker fills in the sender anyway,
2232 overwriting whatever the client filled in.
2234 * sd-event gained a new pseudo-handle that may be specified on all API
2235 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
2236 used this refers to the default event loop object of the calling
2237 thread. Note however that this does not implicitly allocate one —
2238 which has to be done prior by using sd_event_default(). Similarly
2239 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
2240 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
2241 to the default bus of the specified type of the calling thread. Here
2242 too this does not implicitly allocate bus connection objects, this
2243 has to be done prior with sd_bus_default() and friends.
2245 * sd-event gained a new call pair
2246 sd_event_source_{get|set}_io_fd_own(). This may be used to request
2247 automatic closure of the file descriptor an IO event source watches
2248 when the event source is destroyed.
2250 * systemd-networkd gained support for natively configuring WireGuard
2253 * In previous versions systemd synthesized user records both for the
2254 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
2255 internally. In order to simplify distribution-wide renames of the
2256 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
2257 new transitional flag file has been added: if
2258 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
2259 user and group record within the systemd codebase is disabled.
2261 * systemd-notify gained a new --uid= option for selecting the source
2262 user/UID to use for notification messages sent to the service
2265 * journalctl gained a new --grep= option to list only entries in which
2266 the message matches a certain pattern. By default matching is case
2267 insensitive if the pattern is lowercase, and case sensitive
2268 otherwise. Option --case-sensitive=yes|no can be used to override
2269 this an specify case sensitivity or case insensitivity.
2271 * There's now a "systemd-analyze service-watchdogs" command for printing
2272 the current state of the service runtime watchdog, and optionally
2273 enabling or disabling the per-service watchdogs system-wide if given a
2274 boolean argument (i.e. the concept you configure in WatchdogSec=), for
2275 debugging purposes. There's also a kernel command line option
2276 systemd.service_watchdogs= for controlling the same.
2278 * Two new "log-level" and "log-target" options for systemd-analyze were
2279 added that merge the now deprecated get-log-level, set-log-level and
2280 get-log-target, set-log-target pairs. The deprecated options are still
2281 understood for backwards compatibility. The two new options print the
2282 current value when no arguments are given, and set them when a
2283 level/target is given as an argument.
2285 * sysusers.d's "u" lines now optionally accept both a UID and a GID
2286 specification, separated by a ":" character, in order to create users
2287 where UID and GID do not match.
2289 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
2290 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
2291 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
2292 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
2293 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
2294 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
2295 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
2296 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
2297 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
2298 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
2299 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
2300 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
2301 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
2302 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
2303 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
2304 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
2305 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
2306 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
2307 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
2308 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
2315 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
2316 in v235 has been extended to also set the dummy.ko module option
2317 numdummies=0, preventing the kernel from automatically creating
2318 dummy0. All dummy interfaces must now be explicitly created.
2320 * Unknown '%' specifiers in configuration files are now rejected. This
2321 applies to units and tmpfiles.d configuration. Any percent characters
2322 that are followed by a letter or digit that are not supposed to be
2323 interpreted as the beginning of a specifier should be escaped by
2324 doubling ("%%"). (So "size=5%" is still accepted, as well as
2325 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
2326 valid specifiers today.)
2328 * systemd-resolved now maintains a new dynamic
2329 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
2330 recommended to make /etc/resolv.conf a symlink to it. This file
2331 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
2332 includes dynamically acquired search domains, achieving more correct
2333 DNS resolution by software that bypasses local DNS APIs such as NSS.
2335 * The "uaccess" udev tag has been dropped from /dev/kvm and
2336 /dev/dri/renderD*. These devices now have the 0666 permissions by
2337 default (but this may be changed at build-time). /dev/dri/renderD*
2338 will now be owned by the "render" group along with /dev/kfd.
2340 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
2341 systemd-journal-gatewayd.service and
2342 systemd-journal-upload.service. This means "nss-systemd" must be
2343 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
2344 services are resolved properly.
2346 * In /etc/fstab two new mount options are now understood:
2347 x-systemd.makefs and x-systemd.growfs. The former has the effect that
2348 the configured file system is formatted before it is mounted, the
2349 latter that the file system is resized to the full block device size
2350 after it is mounted (i.e. if the file system is smaller than the
2351 partition it resides on, it's grown). This is similar to the fsck
2352 logic in /etc/fstab, and pulls in systemd-makefs@.service and
2353 systemd-growfs@.service as necessary, similar to
2354 systemd-fsck@.service. Resizing is currently only supported on ext4
2357 * In systemd-networkd, the IPv6 RA logic now optionally may announce
2358 DNS server and domain information.
2360 * Support for the LUKS2 on-disk format for encrypted partitions has
2361 been added. This requires libcryptsetup2 during compilation and
2364 * The systemd --user instance will now signal "readiness" when its
2365 basic.target unit has been reached, instead of when the run queue ran
2366 empty for the first time.
2368 * Tmpfiles.d with user configuration are now also supported.
2369 systemd-tmpfiles gained a new --user switch, and snippets placed in
2370 ~/.config/user-tmpfiles.d/ and corresponding directories will be
2371 executed by systemd-tmpfiles --user running in the new
2372 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
2373 running in the user session.
2375 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
2376 %S resolves to the top-level state directory (/var/lib for the system
2377 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
2378 top-level cache directory (/var/cache for the system instance,
2379 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
2380 logs directory (/var/log for the system instance,
2381 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
2382 existing %t specifier, that resolves to the top-level runtime
2383 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
2386 * journalctl learnt a new parameter --output-fields= for limiting the
2387 set of journal fields to output in verbose and JSON output modes.
2389 * systemd-timesyncd's configuration file gained a new option
2390 RootDistanceMaxSec= for setting the maximum root distance of servers
2391 it'll use, as well as the new options PollIntervalMinSec= and
2392 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
2394 * bootctl gained a new command "list" for listing all available boot
2395 menu items on systems that follow the boot loader specification.
2397 * systemctl gained a new --dry-run switch that shows what would be done
2398 instead of doing it, and is currently supported by the shutdown and
2401 * ConditionSecurity= can now detect the TOMOYO security module.
2403 * Unit file [Install] sections are now also respected in unit drop-in
2404 files. This is intended to be used by drop-ins under /usr/lib/.
2406 * systemd-firstboot may now also set the initial keyboard mapping.
2408 * Udev "changed" events for devices which are exposed as systemd
2409 .device units are now propagated to units specified in
2410 ReloadPropagatedFrom= as reload requests.
2412 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
2413 unit template name (i.e. a name in the form of 'foobar@.service',
2414 without the instance component between the '@' and - the '.'), then
2415 the escaped sysfs path of the device is automatically used as the
2418 * SystemCallFilter= in unit files has been extended so that an "errno"
2419 can be specified individually for each system call. Example:
2420 SystemCallFilter=~uname:EILSEQ.
2422 * The cgroup delegation logic has been substantially updated. Delegate=
2423 now optionally takes a list of controllers (instead of a boolean, as
2424 before), which lists the controllers to delegate at least.
2426 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
2428 * A new LogLevelMax= setting configures the maximum log level any
2429 process of the service may log at (i.e. anything with a lesser
2430 priority than what is specified is automatically dropped). A new
2431 LogExtraFields= setting allows configuration of additional journal
2432 fields to attach to all log records generated by any of the unit's
2435 * New StandardInputData= and StandardInputText= settings along with the
2436 new option StandardInput=data may be used to configure textual or
2437 binary data that shall be passed to the executed service process via
2438 standard input, encoded in-line in the unit file.
2440 * StandardInput=, StandardOutput= and StandardError= may now be used to
2441 connect stdin/stdout/stderr of executed processes directly with a
2442 file or AF_UNIX socket in the file system, using the new "file:" option.
2444 * A new unit file option CollectMode= has been added, that allows
2445 tweaking the garbage collection logic for units. It may be used to
2446 tell systemd to garbage collect units that have failed automatically
2447 (normally it only GCs units that exited successfully). systemd-run
2448 and systemd-mount expose this new functionality with a new -G option.
2450 * "machinectl bind" may now be used to bind mount non-directories
2451 (i.e. regularfiles, devices, fifos, sockets).
2453 * systemd-analyze gained a new verb "calendar" for validating and
2454 testing calendar time specifications to use for OnCalendar= in timer
2455 units. Besides validating the expression it will calculate the next
2456 time the specified expression would elapse.
2458 * In addition to the pre-existing FailureAction= unit file setting
2459 there's now SuccessAction=, for configuring a shutdown action to
2460 execute when a unit completes successfully. This is useful in
2461 particular inside containers that shall terminate after some workload
2462 has been completed. Also, both options are now supported for all unit
2463 types, not just services.
2465 * networkds's IP rule support gained two new options
2466 IncomingInterface= and OutgoingInterface= for configuring the incoming
2467 and outgoing interfaces of configured rules. systemd-networkd also
2468 gained support for "vxcan" network devices.
2470 * networkd gained a new setting RequiredForOnline=, taking a
2471 boolean. If set, systemd-wait-online will take it into consideration
2472 when determining that the system is up, otherwise it will ignore the
2473 interface for this purpose.
2475 * The sd_notify() protocol gained support for a new operation: with
2476 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
2477 store again, ahead of POLLHUP or POLLERR when they are removed
2480 * A new document doc/UIDS-GIDS.md has been added to the source tree,
2481 that documents the UID/GID range and assignment assumptions and
2482 requirements of systemd.
2484 * The watchdog device PID 1 will ping may now be configured through the
2485 WatchdogDevice= configuration file setting, or by setting the
2486 systemd.watchdog_service= kernel commandline option.
2488 * systemd-resolved's gained support for registering DNS-SD services on
2489 the local network using MulticastDNS. Services may either be
2490 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
2491 the same dir below /run, /usr/lib), or through its D-Bus API.
2493 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
2494 extend the effective start, runtime, and stop time. The service must
2495 continue to send EXTEND_TIMEOUT_USEC within the period specified to
2496 prevent the service manager from making the service as timedout.
2498 * systemd-resolved's DNSSEC support gained support for RFC 8080
2499 (Ed25519 keys and signatures).
2501 * The systemd-resolve command line tool gained a new set of options
2502 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
2503 --set-nta= and --revert to configure per-interface DNS configuration
2504 dynamically during runtime. It's useful for pushing DNS information
2505 into systemd-resolved from DNS hook scripts that various interface
2506 managing software supports (such as pppd).
2508 * systemd-nspawn gained a new --network-namespace-path= command line
2509 option, which may be used to make a container join an existing
2510 network namespace, by specifying a path to a "netns" file.
2512 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
2513 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
2514 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
2515 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
2516 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
2517 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
2518 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
2519 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
2520 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
2521 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
2522 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
2523 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
2524 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
2525 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
2526 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
2527 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
2528 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
2529 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
2530 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
2531 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
2532 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
2533 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
2534 Jędrzejewski-Szmek, Zeal Jagannatha
2536 — Berlin, 2017-12-14
2540 * INCOMPATIBILITY: systemd-logind.service and other long-running
2541 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
2542 communication with the outside. This generally improves security of
2543 the system, and is in almost all cases a safe and good choice, as
2544 these services do not and should not provide any network-facing
2545 functionality. However, systemd-logind uses the glibc NSS API to
2546 query the user database. This creates problems on systems where NSS
2547 is set up to directly consult network services for user database
2548 lookups. In particular, this creates incompatibilities with the
2549 "nss-nis" module, which attempts to directly contact the NIS/YP
2550 network servers it is configured for, and will now consistently
2551 fail. In such cases, it is possible to turn off IP sandboxing for
2552 systemd-logind.service (set IPAddressDeny= in its [Service] section
2553 to the empty string, via a .d/ unit file drop-in). Downstream
2554 distributions might want to update their nss-nis packaging to include
2555 such a drop-in snippet, accordingly, to hide this incompatibility
2556 from the user. Another option is to make use of glibc's nscd service
2557 to proxy such network requests through a privilege-separated, minimal
2558 local caching daemon, or to switch to more modern technologies such
2559 sssd, whose NSS hook-ups generally do not involve direct network
2560 access. In general, we think it's definitely time to question the
2561 implementation choices of nss-nis, i.e. whether it's a good idea
2562 today to embed a network-facing loadable module into all local
2563 processes that need to query the user database, including the most
2564 trivial and benign ones, such as "ls". For more details about
2565 IPAddressDeny= see below.
2567 * A new modprobe.d drop-in is now shipped by default that sets the
2568 bonding module option max_bonds=0. This overrides the kernel default,
2569 to avoid conflicts and ambiguity as to whether or not bond0 should be
2570 managed by systemd-networkd or not. This resolves multiple issues
2571 with bond0 properties not being applied, when bond0 is configured
2572 with systemd-networkd. Distributors may choose to not package this,
2573 however in that case users will be prevented from correctly managing
2574 bond0 interface using systemd-networkd.
2576 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
2577 which print the logging level and target of the system manager. They
2578 complement the existing "set-log-level" and "set-log-target" verbs
2579 used to change those values.
2581 * journald.conf gained a new boolean setting ReadKMsg= which defaults
2582 to on. If turned off kernel log messages will not be read by
2583 systemd-journald or included in the logs. It also gained a new
2584 setting LineMax= for configuring the maximum line length in
2585 STDOUT/STDERR log streams. The new default for this value is 48K, up
2586 from the previous hardcoded 2048.
2588 * A new unit setting RuntimeDirectoryPreserve= has been added, which
2589 allows more detailed control of what to do with a runtime directory
2590 configured with RuntimeDirectory= (i.e. a directory below /run or
2591 $XDG_RUNTIME_DIR) after a unit is stopped.
2593 * The RuntimeDirectory= setting for units gained support for creating
2594 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
2595 one top-level directory.
2597 * Units gained new options StateDirectory=, CacheDirectory=,
2598 LogsDirectory= and ConfigurationDirectory= which are closely related
2599 to RuntimeDirectory= but manage per-service directories below
2600 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
2601 possible to write unit files which when activated automatically gain
2602 properly owned service specific directories in these locations, thus
2603 making unit files self-contained and increasing compatibility with
2604 stateless systems and factory reset where /etc or /var are
2605 unpopulated at boot. Matching these new settings there's also
2606 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
2607 ConfigurationDirectoryMode= for configuring the access mode of these
2608 directories. These settings are particularly useful in combination
2609 with DynamicUser=yes as they provide secure, properly-owned,
2610 writable, and stateful locations for storage, excluded from the
2611 sandbox that such services live in otherwise.
2613 * Automake support has been removed from this release. systemd is now
2616 * systemd-journald will now aggressively cache client metadata during
2617 runtime, speeding up log write performance under pressure. This comes
2618 at a small price though: as much of the metadata is read
2619 asynchronously from /proc/ (and isn't implicitly attached to log
2620 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
2621 metadata stored alongside a log entry might be slightly
2622 out-of-date. Previously it could only be slightly newer than the log
2623 message. The time window is small however, and given that the kernel
2624 is unlikely to be improved anytime soon in this regard, this appears
2627 * nss-myhostname/systemd-resolved will now by default synthesize an
2628 A/AAAA resource record for the "_gateway" hostname, pointing to the
2629 current default IP gateway. Previously it did that for the "gateway"
2630 name, hampering adoption, as some distributions wanted to leave that
2631 host name open for local use. The old behaviour may still be
2632 requested at build time.
2634 * systemd-networkd's [Address] section in .network files gained a new
2635 Scope= setting for configuring the IP address scope. The [Network]
2636 section gained a new boolean setting ConfigureWithoutCarrier= that
2637 tells systemd-networkd to ignore link sensing when configuring the
2638 device. The [DHCP] section gained a new Anonymize= boolean option for
2639 turning on a number of options suggested in RFC 7844. A new
2640 [RoutingPolicyRule] section has been added for configuring the IP
2641 routing policy. The [Route] section has gained support for a new
2642 Type= setting which permits configuring
2643 blackhole/unreachable/prohibit routes.
2645 * The [VRF] section in .netdev files gained a new Table= setting for
2646 configuring the routing table to use. The [Tunnel] section gained a
2647 new Independent= boolean field for configuring tunnels independent of
2648 an underlying network interface. The [Bridge] section gained a new
2649 GroupForwardMask= option for configuration of propagation of link
2650 local frames between bridge ports.
2652 * The WakeOnLan= setting in .link files gained support for a number of
2653 new modes. A new TCP6SegmentationOffload= setting has been added for
2654 configuring TCP/IPv6 hardware segmentation offload.
2656 * The IPv6 RA sender implementation may now optionally send out RDNSS
2657 and RDNSSL records to supply DNS configuration to peers.
2659 * systemd-nspawn gained support for a new --system-call-filter= command
2660 line option for adding and removing entries in the default system
2661 call filter it applies. Moreover systemd-nspawn has been changed to
2662 implement a system call whitelist instead of a blacklist.
2664 * systemd-run gained support for a new --pipe command line option. If
2665 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
2666 are directly passed on to the activated transient service
2667 executable. This allows invoking arbitrary processes as systemd
2668 services (for example to take benefit of dependency management,
2669 accounting management, resource management or log management that is
2670 done automatically for services) — while still allowing them to be
2671 integrated in a classic UNIX shell pipeline.
2673 * When a service sends RELOAD=1 via sd_notify() and reload propagation
2674 using ReloadPropagationTo= is configured, a reload is now propagated
2675 to configured units. (Previously this was only done on explicitly
2676 requested reloads, using "systemctl reload" or an equivalent
2679 * For each service unit a restart counter is now kept: it is increased
2680 each time the service is restarted due to Restart=, and may be
2681 queried using "systemctl show -p NRestarts …".
2683 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
2684 @signal and @timer have been added, for usage with SystemCallFilter=
2685 in unit files and the new --system-call-filter= command line option
2686 of systemd-nspawn (see above).
2688 * ExecStart= lines in unit files gained two new modifiers: when a
2689 command line is prefixed with "!" the command will be executed as
2690 configured, except for the credentials applied by
2691 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
2692 "+", but does still apply namespacing options unlike "+". There's
2693 also "!!" now, which is mostly identical, but becomes a NOP on
2694 systems that support ambient capabilities. This is useful to write
2695 unit files that work with ambient capabilities where possible but
2696 automatically fall back to traditional privilege dropping mechanisms
2697 on systems where this is not supported.
2699 * ListenNetlink= settings in socket units now support RDMA netlink
2702 * A new unit file setting LockPersonality= has been added which permits
2703 locking down the chosen execution domain ("personality") of a service
2706 * A new special target "getty-pre.target" has been added, which is
2707 ordered before all text logins, and may be used to order services
2708 before textual logins acquire access to the console.
2710 * systemd will now attempt to load the virtio-rng.ko kernel module very
2711 early on if a VM environment supporting this is detected. This should
2712 improve entropy during early boot in virtualized environments.
2714 * A _netdev option is now supported in /etc/crypttab that operates in a
2715 similar way as the same option in /etc/fstab: it permits configuring
2716 encrypted devices that need to be ordered after the network is up.
2717 Following this logic, two new special targets
2718 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
2719 added that are to cryptsetup.target what remote-fs.target and
2720 remote-fs-pre.target are to local-fs.target.
2722 * Service units gained a new UnsetEnvironment= setting which permits
2723 unsetting specific environment variables for services that are
2724 normally passed to it (for example in order to mask out locale
2725 settings for specific services that can't deal with it).
2727 * Units acquired a new boolean option IPAccounting=. When turned on, IP
2728 traffic accounting (packet count as well as byte count) is done for
2729 the service, and shown as part of "systemctl status" or "systemd-run
2732 * Service units acquired two new options IPAddressAllow= and
2733 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
2734 for configuring a simple IP access control list for all sockets of
2735 the unit. These options are available also on .slice and .socket
2736 units, permitting flexible access list configuration for individual
2737 services as well as groups of services (as defined by a slice unit),
2738 including system-wide. Note that IP ACLs configured this way are
2739 enforced on every single IPv4 and IPv6 socket created by any process
2740 of the service unit, and apply to ingress as well as egress traffic.
2742 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
2743 structured log message is generated each time the unit is stopped,
2744 containing information about the consumed resources of this
2747 * A new setting KeyringMode= has been added to unit files, which may be
2748 used to control how the kernel keyring is set up for executed
2751 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
2752 "systemctl kexec" and "systemctl exit" are now always asynchronous in
2753 behaviour (that is: these commands return immediately after the
2754 operation was enqueued instead of waiting for the operation to
2755 complete). Previously, "systemctl poweroff" and "systemctl reboot"
2756 were asynchronous on systems using systemd-logind (i.e. almost
2757 always, and like they were on sysvinit), and the other three commands
2758 were unconditionally synchronous. With this release this is cleaned
2759 up, and callers will see the same asynchronous behaviour on all
2760 systems for all five operations.
2762 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
2765 * .timer units now accept calendar specifications in other timezones
2766 than UTC or the local timezone.
2768 * The tmpfiles snippet var.conf has been changed to create
2769 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
2770 the "utmp" group already, and it appears to be generally understood
2771 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
2772 databases. Previously this was implemented correctly for all these
2773 databases excepts btmp, which has been opened up like this now
2774 too. Note that while the other databases are world-readable
2775 (i.e. 0644), btmp is not and remains more restrictive.
2777 * The systemd-resolve tool gained a new --reset-server-features
2778 switch. When invoked like this systemd-resolved will forget
2779 everything it learnt about the features supported by the configured
2780 upstream DNS servers, and restarts the feature probing logic on the
2781 next resolver look-up for them at the highest feature level
2784 * The status dump systemd-resolved sends to the logs upon receiving
2785 SIGUSR1 now also includes information about all DNS servers it is
2786 configured to use, and the features levels it probed for them.
2788 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
2789 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
2790 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
2791 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
2792 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
2793 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
2794 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
2795 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
2796 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
2797 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
2798 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
2799 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
2800 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
2801 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
2802 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
2803 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
2804 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
2805 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
2806 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
2807 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2809 — Berlin, 2017-10-06
2813 * Meson is now supported as build system in addition to Automake. It is
2814 our plan to remove Automake in one of our next releases, so that
2815 Meson becomes our exclusive build system. Hence, please start using
2816 the Meson build system in your downstream packaging. There's plenty
2817 of documentation around how to use Meson, the extremely brief
2820 ./autogen.sh && ./configure && make && sudo make install
2824 meson build && ninja -C build && sudo ninja -C build install
2826 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
2827 which permits configuring a timeout on the time a job is
2828 running. This is particularly useful for setting timeouts on jobs for
2831 * Unit files gained two new options ConditionUser= and ConditionGroup=
2832 for conditionalizing units based on the identity of the user/group
2833 running a systemd user instance.
2835 * systemd-networkd now understands a new FlowLabel= setting in the
2836 [VXLAN] section of .network files, as well as a Priority= in
2837 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
2838 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
2839 gained support for configuration of GENEVE links, and IPv6 address
2840 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
2842 * .link files now understand a new Port= setting.
2844 * systemd-networkd's DHCP support gained support for DHCP option 119
2845 (domain search list).
2847 * systemd-networkd gained support for serving IPv6 address ranges using
2848 the Router Advertisement protocol. The new .network configuration
2849 section [IPv6Prefix] may be used to configure the ranges to
2850 serve. This is implemented based on a new, minimal, native server
2851 implementation of RA.
2853 * journalctl's --output= switch gained support for a new parameter
2854 "short-iso-precise" for a mode where timestamps are shown as precise
2857 * systemd-udevd's "net_id" builtin may now generate stable network
2858 interface names from IBM PowerVM VIO devices as well as ACPI platform
2861 * MulticastDNS support in systemd-resolved may now be explicitly
2862 enabled/disabled using the new MulticastDNS= configuration file
2865 * systemd-resolved may now optionally use libidn2 instead of the libidn
2866 for processing internationalized domain names. Support for libidn2
2867 should be considered experimental and should not be enabled by
2870 * "machinectl pull-tar" and related call may now do verification of
2871 downloaded images using SUSE-style .sha256 checksum files in addition
2872 to the already existing support for validating using Ubuntu-style
2875 * sd-bus gained support for a new sd_bus_message_appendv() call which
2876 is va_list equivalent of sd_bus_message_append().
2878 * sd-boot gained support for validating images using SHIM/MOK.
2880 * The SMACK code learnt support for "onlycap".
2882 * systemd-mount --umount is now much smarter in figuring out how to
2883 properly unmount a device given its mount or device path.
2885 * The code to call libnss_dns as a fallback from libnss_resolve when
2886 the communication with systemd-resolved fails was removed. This
2887 fallback was redundant and interfered with the [!UNAVAIL=return]
2888 suffix. See nss-resolve(8) for the recommended configuration.
2890 * systemd-logind may now be restarted without losing state. It stores
2891 the file descriptors for devices it manages in the system manager
2892 using the FDSTORE= mechanism. Please note that further changes in
2893 other components may be required to make use of this (for example
2894 Xorg has code to listen for stops of systemd-logind and terminate
2895 itself when logind is stopped or restarted, in order to avoid using
2896 stale file descriptors for graphical devices, which is now
2897 counterproductive and must be reverted in order for restarts of
2898 systemd-logind to be safe. See
2899 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
2901 * All kernel-install plugins are called with the environment variable
2902 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
2903 /etc/machine-id. If the machine ID could not be determined,
2904 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
2905 anything in the entry directory (passed as the second argument) if
2906 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
2907 temporary directory is passed as the entry directory and removed
2908 after all the plugins exit.
2910 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
2911 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
2912 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
2913 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
2914 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
2915 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
2916 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
2917 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
2918 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
2919 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
2920 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
2921 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
2922 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
2923 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
2924 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
2925 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
2926 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
2927 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
2928 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
2929 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
2930 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
2931 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
2932 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
2933 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
2934 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
2935 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
2936 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
2937 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
2938 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
2941 — Berlin, 2017-07-12
2945 * The "hybrid" control group mode has been modified to improve
2946 compatibility with "legacy" cgroups-v1 setups. Specifically, the
2947 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
2948 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
2949 cgroups-v1 hierarchy), the only externally visible change being that
2950 the cgroups-v2 hierarchy is also mounted, to
2951 /sys/fs/cgroup/unified. This should provide a large degree of
2952 compatibility with "legacy" cgroups-v1, while taking benefit of the
2953 better management capabilities of cgroups-v2.
2955 * The default control group setup mode may be selected both a boot-time
2956 via a set of kernel command line parameters (specifically:
2957 systemd.unified_cgroup_hierarchy= and
2958 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
2959 default selected on the configure command line
2960 (--with-default-hierarchy=). The upstream default is "hybrid"
2961 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
2962 this will change in a future systemd version to be "unified" (pure
2963 cgroups-v2 mode). The third option for the compile time option is
2964 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
2965 distributions to default to "hybrid" mode for release distributions,
2966 starting with v233. We recommend "unified" for development
2967 distributions (specifically: distributions such as Fedora's rawhide)
2968 as that's where things are headed in the long run. Use "legacy" for
2969 greatest stability and compatibility only.
2971 * Note one current limitation of "unified" and "hybrid" control group
2972 setup modes: the kernel currently does not permit the systemd --user
2973 instance (i.e. unprivileged code) to migrate processes between two
2974 disconnected cgroup subtrees, even if both are managed and owned by
2975 the user. This effectively means "systemd-run --user --scope" doesn't
2976 work when invoked from outside of any "systemd --user" service or
2977 scope. Specifically, it is not supported from session scopes. We are
2978 working on fixing this in a future systemd version. (See #3388 for
2979 further details about this.)
2981 * DBus policy files are now installed into /usr rather than /etc. Make
2982 sure your system has dbus >= 1.9.18 running before upgrading to this
2983 version, or override the install path with --with-dbuspolicydir= .
2985 * All python scripts shipped with systemd (specifically: the various
2986 tests written in Python) now require Python 3.
2988 * systemd unit tests can now run standalone (without the source or
2989 build directories), and can be installed into /usr/lib/systemd/tests/
2990 with 'make install-tests'.
2992 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
2993 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
2996 * Support for the %c, %r, %R specifiers in unit files has been
2997 removed. Specifiers are not supposed to be dependent on configuration
2998 in the unit file itself (so that they resolve the same regardless
2999 where used in the unit files), but these specifiers were influenced
3000 by the Slice= option.
3002 * The shell invoked by debug-shell.service now defaults to /bin/sh in
3003 all cases. If distributions want to use a different shell for this
3004 purpose (for example Fedora's /sbin/sushell) they need to specify
3005 this explicitly at configure time using --with-debug-shell=.
3007 * The confirmation spawn prompt has been reworked to offer the
3010 (c)ontinue, proceed without asking anymore
3011 (D)ump, show the state of the unit
3012 (f)ail, don't execute the command and pretend it failed
3014 (i)nfo, show a short summary of the unit
3015 (j)obs, show jobs that are in progress
3016 (s)kip, don't execute the command and pretend it succeeded
3017 (y)es, execute the command
3019 The 'n' choice for the confirmation spawn prompt has been removed,
3020 because its meaning was confusing.
3022 The prompt may now also be redirected to an alternative console by
3023 specifying the console as parameter to systemd.confirm_spawn=.
3025 * Services of Type=notify require a READY=1 notification to be sent
3026 during startup. If no such message is sent, the service now fails,
3027 even if the main process exited with a successful exit code.
3029 * Services that fail to start up correctly now always have their
3030 ExecStopPost= commands executed. Previously, they'd enter "failed"
3031 state directly, without executing these commands.
3033 * The option MulticastDNS= of network configuration files has acquired
3034 an actual implementation. With MulticastDNS=yes a host can resolve
3035 names of remote hosts and reply to mDNS A and AAAA requests.
3037 * When units are about to be started an additional check is now done to
3038 ensure that all dependencies of type BindsTo= (when used in
3039 combination with After=) have been started.
3041 * systemd-analyze gained a new verb "syscall-filter" which shows which
3042 system call groups are defined for the SystemCallFilter= unit file
3043 setting, and which system calls they contain.
3045 * A new system call filter group "@filesystem" has been added,
3046 consisting of various file system related system calls. Group
3047 "@reboot" has been added, covering reboot, kexec and shutdown related
3048 calls. Finally, group "@swap" has been added covering swap
3049 configuration related calls.
3051 * A new unit file option RestrictNamespaces= has been added that may be
3052 used to restrict access to the various process namespace types the
3053 Linux kernel provides. Specifically, it may be used to take away the
3054 right for a service unit to create additional file system, network,
3055 user, and other namespaces. This sandboxing option is particularly
3056 relevant due to the high amount of recently discovered namespacing
3057 related vulnerabilities in the kernel.
3059 * systemd-udev's .link files gained support for a new AutoNegotiation=
3060 setting for configuring Ethernet auto-negotiation.
3062 * systemd-networkd's .network files gained support for a new
3063 ListenPort= setting in the [DHCP] section to explicitly configure the
3064 UDP client port the DHCP client shall listen on.
3066 * .network files gained a new Unmanaged= boolean setting for explicitly
3067 excluding one or more interfaces from management by systemd-networkd.
3069 * The systemd-networkd ProxyARP= option has been renamed to
3070 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
3071 renamed to ReduceARPProxy=. The old names continue to be available
3074 * systemd-networkd gained support for configuring IPv6 Proxy NDP
3075 addresses via the new IPv6ProxyNDPAddress= .network file setting.
3077 * systemd-networkd's bonding device support gained support for two new
3078 configuration options ActiveSlave= and PrimarySlave=.
3080 * The various options in the [Match] section of .network files gained
3081 support for negative matching.
3083 * New systemd-specific mount options are now understood in /etc/fstab:
3085 x-systemd.mount-timeout= may be used to configure the maximum
3086 permitted runtime of the mount command.
3088 x-systemd.device-bound may be set to bind a mount point to its
3089 backing device unit, in order to automatically remove a mount point
3090 if its backing device is unplugged. This option may also be
3091 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
3092 on the block device, which is now automatically set for all CDROM
3093 drives, so that mounted CDs are automatically unmounted when they are
3094 removed from the drive.
3096 x-systemd.after= and x-systemd.before= may be used to explicitly
3097 order a mount after or before another unit or mount point.
3099 * Enqueued start jobs for device units are now automatically garbage
3100 collected if there are no jobs waiting for them anymore.
3102 * systemctl list-jobs gained two new switches: with --after, for every
3103 queued job the jobs it's waiting for are shown; with --before the
3104 jobs which it's blocking are shown.
3106 * systemd-nspawn gained support for ephemeral boots from disk images
3107 (or in other words: --ephemeral and --image= may now be
3108 combined). Moreover, ephemeral boots are now supported for normal
3109 directories, even if the backing file system is not btrfs. Of course,
3110 if the file system does not support file system snapshots or
3111 reflinks, the initial copy operation will be relatively expensive, but
3112 this should still be suitable for many use cases.
3114 * Calendar time specifications in .timer units now support
3115 specifications relative to the end of a month by using "~" instead of
3116 "-" as separator between month and day. For example, "*-02~03" means
3117 "the third last day in February". In addition a new syntax for
3118 repeated events has been added using the "/" character. For example,
3119 "9..17/2:00" means "every two hours from 9am to 5pm".
3121 * systemd-socket-proxyd gained a new parameter --connections-max= for
3122 configuring the maximum number of concurrent connections.
3124 * sd-id128 gained a new API for generating unique IDs for the host in a
3125 way that does not leak the machine ID. Specifically,
3126 sd_id128_get_machine_app_specific() derives an ID based on the
3127 machine ID a in well-defined, non-reversible, stable way. This is
3128 useful whenever an identifier for the host is needed but where the
3129 identifier shall not be useful to identify the system beyond the
3130 scope of the application itself. (Internally this uses HMAC-SHA256 as
3131 keyed hash function using the machine ID as input.)
3133 * NotifyAccess= gained a new supported value "exec". When set
3134 notifications are accepted from all processes systemd itself invoked,
3135 including all control processes.
3137 * .nspawn files gained support for defining overlay mounts using the
3138 Overlay= and OverlayReadOnly= options. Previously this functionality
3139 was only available on the systemd-nspawn command line.
3141 * systemd-nspawn's --bind= and --overlay= options gained support for
3142 bind/overlay mounts whose source lies within the container tree by
3143 prefixing the source path with "+".
3145 * systemd-nspawn's --bind= and --overlay= options gained support for
3146 automatically allocating a temporary source directory in /var/tmp
3147 that is removed when the container dies. Specifically, if the source
3148 directory is specified as empty string this mechanism is selected. An
3149 example usage is --overlay=+/var::/var, which creates an overlay
3150 mount based on the original /var contained in the image, overlayed
3151 with a temporary directory in the host's /var/tmp. This way changes
3152 to /var are automatically flushed when the container shuts down.
3154 * systemd-nspawn --image= option does now permit raw file system block
3155 devices (in addition to images containing partition tables, as
3158 * The disk image dissection logic in systemd-nspawn gained support for
3159 automatically setting up LUKS encrypted as well as Verity protected
3160 partitions. When a container is booted from an encrypted image the
3161 passphrase is queried at start-up time. When a container with Verity
3162 data is started, the root hash is search in a ".roothash" file
3163 accompanying the disk image (alternatively, pass the root hash via
3164 the new --root-hash= command line option).
3166 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
3167 be used to dissect disk images the same way as systemd-nspawn does
3168 it, following the Bootable Partition Specification. It may even be
3169 used to mount disk images with complex partition setups (including
3170 LUKS and Verity partitions) to a local host directory, in order to
3171 inspect them. This tool is not considered public API (yet), and is
3172 thus not installed into /usr/bin. Please do not rely on its
3173 existence, since it might go away or be changed in later systemd
3176 * A new generator "systemd-verity-generator" has been added, similar in
3177 style to "systemd-cryptsetup-generator", permitting automatic setup of
3178 Verity root partitions when systemd boots up. In order to make use of
3179 this your partition setup should follow the Discoverable Partitions
3180 Specification, and the GPT partition ID of the root file system
3181 partition should be identical to the upper 128bit of the Verity root
3182 hash. The GPT partition ID of the Verity partition protecting it
3183 should be the lower 128bit of the Verity root hash. If the partition
3184 image follows this model it is sufficient to specify a single
3185 "roothash=" kernel command line argument to both configure which root
3186 image and verity partition to use as well as the root hash for
3187 it. Note that systemd-nspawn's Verity support follows the same
3188 semantics, meaning that disk images with proper Verity data in place
3189 may be booted in containers with systemd-nspawn as well as on
3190 physical systems via the verity generator. Also note that the "mkosi"
3191 tool available at https://github.com/systemd/mkosi has been updated
3192 to generate Verity protected disk images following this scheme. In
3193 fact, it has been updated to generate disk images that optionally
3194 implement a complete UEFI SecureBoot trust chain, involving a signed
3195 kernel and initrd image that incorporates such a root hash as well as
3196 a Verity-enabled root partition.
3198 * The hardware database (hwdb) udev supports has been updated to carry
3199 accelerometer quirks.
3201 * All system services are now run with a fresh kernel keyring set up
3202 for them. The invocation ID is stored by default in it, thus
3203 providing a safe, non-overridable way to determine the invocation
3206 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
3207 options for bind mounting arbitrary paths in a service-specific
3208 way. When these options are used, arbitrary host or service files and
3209 directories may be mounted to arbitrary locations in the service's
3212 * Documentation has been added that lists all of systemd's low-level
3213 environment variables:
3215 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
3217 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
3218 whether a specific socket file descriptor matches a specified socket
3221 * systemd-firstboot has been updated to check for the
3222 systemd.firstboot= kernel command line option. It accepts a boolean
3223 and when set to false the first boot questions are skipped.
3225 * systemd-fstab-generator has been updated to check for the
3226 systemd.volatile= kernel command line option, which either takes an
3227 optional boolean parameter or the special value "state". If used the
3228 system may be booted in a "volatile" boot mode. Specifically,
3229 "systemd.volatile" is used, the root directory will be mounted as
3230 tmpfs, and only /usr is mounted from the actual root file system. If
3231 "systemd.volatile=state" is used, the root directory will be mounted
3232 as usual, but /var is mounted as tmpfs. This concept provides similar
3233 functionality as systemd-nspawn's --volatile= option, but provides it
3234 on physical boots. Use this option for implementing stateless
3235 systems, or testing systems with all state and/or configuration reset
3236 to the defaults. (Note though that many distributions are not
3237 prepared to boot up without a populated /etc or /var, though.)
3239 * systemd-gpt-auto-generator gained support for LUKS encrypted root
3240 partitions. Previously it only supported LUKS encrypted partitions
3241 for all other uses, except for the root partition itself.
3243 * Socket units gained support for listening on AF_VSOCK sockets for
3244 communication in virtualized QEMU environments.
3246 * The "configure" script gained a new option --with-fallback-hostname=
3247 for specifying the fallback hostname to use if none is configured in
3248 /etc/hostname. For example, by specifying
3249 --with-fallback-hostname=fedora it is possible to default to a
3250 hostname of "fedora" on pristine installations.
3252 * systemd-cgls gained support for a new --unit= switch for listing only
3253 the control groups of a specific unit. Similar --user-unit= has been
3254 added for listing only the control groups of a specific user unit.
3256 * systemd-mount gained a new --umount switch for unmounting a mount or
3257 automount point (and all mount/automount points below it).
3259 * systemd will now refuse full configuration reloads (via systemctl
3260 daemon-reload and related calls) unless at least 16MiB of free space
3261 are available in /run. This is a safety precaution in order to ensure
3262 that generators can safely operate after the reload completed.
3264 * A new unit file option RootImage= has been added, which has a similar
3265 effect as RootDirectory= but mounts the service's root directory from
3266 a disk image instead of plain directory. This logic reuses the same
3267 image dissection and mount logic that systemd-nspawn already uses,
3268 and hence supports any disk images systemd-nspawn supports, including
3269 those following the Discoverable Partition Specification, as well as
3270 Verity enabled images. This option enables systemd to run system
3271 services directly off disk images acting as resource bundles,
3272 possibly even including full integrity data.
3274 * A new MountAPIVFS= unit file option has been added, taking a boolean
3275 argument. If enabled /proc, /sys and /dev (collectively called the
3276 "API VFS") will be mounted for the service. This is only relevant if
3277 RootDirectory= or RootImage= is used for the service, as these mounts
3278 are of course in place in the host mount namespace anyway.
3280 * systemd-nspawn gained support for a new --pivot-root= switch. If
3281 specified the root directory within the container image is pivoted to
3282 the specified mount point, while the original root disk is moved to a
3283 different place. This option enables booting of ostree images
3284 directly with systemd-nspawn.
3286 * The systemd build scripts will no longer complain if the NTP server
3287 addresses are not changed from the defaults. Google now supports
3288 these NTP servers officially. We still recommend downstreams to
3289 properly register an NTP pool with the NTP pool project though.
3291 * coredumpctl gained a new "--reverse" option for printing the list
3292 of coredumps in reverse order.
3294 * coredumpctl will now show additional information about truncated and
3295 inaccessible coredumps, as well as coredumps that are still being
3296 processed. It also gained a new --quiet switch for suppressing
3297 additional informational message in its output.
3299 * coredumpctl gained support for only showing coredumps newer and/or
3300 older than specific timestamps, using the new --since= and --until=
3301 options, reminiscent of journalctl's options by the same name.
3303 * The systemd-coredump logic has been improved so that it may be reused
3304 to collect backtraces in non-compiled languages, for example in
3305 scripting languages such as Python.
3307 * machinectl will now show the UID shift of local containers, if user
3308 namespacing is enabled for them.
3310 * systemd will now optionally run "environment generator" binaries at
3311 configuration load time. They may be used to add environment
3312 variables to the environment block passed to services invoked. One
3313 user environment generator is shipped by default that sets up
3314 environment variables based on files dropped into /etc/environment.d
3315 and ~/.config/environment.d/.
3317 * systemd-resolved now includes the new, recently published 2017 DNSSEC
3320 * hostnamed has been updated to report a new chassis type of
3321 "convertible" to cover "foldable" laptops that can both act as a
3322 tablet and as a laptop, such as various Lenovo Yoga devices.
3324 Contributions from: Adrián López, Alexander Galanin, Alexander
3325 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
3326 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
3327 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
3328 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
3329 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
3330 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
3331 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
3332 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
3333 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
3334 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
3335 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
3336 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
3337 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
3338 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
3339 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
3340 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
3341 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
3342 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
3343 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
3344 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
3345 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
3346 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
3347 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
3348 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
3349 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
3350 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
3351 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
3354 — Berlin, 2017-03-01
3358 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
3359 RestrictAddressFamilies= enabled. These sandboxing options should
3360 generally be compatible with the various external udev call-out
3361 binaries we are aware of, however there may be exceptions, in
3362 particular when exotic languages for these call-outs are used. In
3363 this case, consider turning off these settings locally.
3365 * The new RemoveIPC= option can be used to remove IPC objects owned by
3366 the user or group of a service when that service exits.
3368 * The new ProtectKernelModules= option can be used to disable explicit
3369 load and unload operations of kernel modules by a service. In
3370 addition access to /usr/lib/modules is removed if this option is set.
3372 * ProtectSystem= option gained a new value "strict", which causes the
3373 whole file system tree with the exception of /dev, /proc, and /sys,
3374 to be remounted read-only for a service.
3376 * The new ProtectKernelTunables= option can be used to disable
3377 modification of configuration files in /sys and /proc by a service.
3378 Various directories and files are remounted read-only, so access is
3379 restricted even if the file permissions would allow it.
3381 * The new ProtectControlGroups= option can be used to disable write
3382 access by a service to /sys/fs/cgroup.
3384 * Various systemd services have been hardened with
3385 ProtectKernelTunables=yes, ProtectControlGroups=yes,
3386 RestrictAddressFamilies=.
3388 * Support for dynamically creating users for the lifetime of a service
3389 has been added. If DynamicUser=yes is specified, user and group IDs
3390 will be allocated from the range 61184..65519 for the lifetime of the
3391 service. They can be resolved using the new nss-systemd.so NSS
3392 module. The module must be enabled in /etc/nsswitch.conf. Services
3393 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
3394 any resources allocated by the service will be cleaned up when the
3395 service exits. They also have ProtectHome=read-only and
3396 ProtectSystem=strict enabled, so they are not able to make any
3397 permanent modifications to the system.
3399 * The nss-systemd module also always resolves root and nobody, making
3400 it possible to have no /etc/passwd or /etc/group files in minimal
3401 container or chroot environments.
3403 * Services may be started with their own user namespace using the new
3404 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
3405 under which the service is running are mapped. All other users are
3408 * Support for the cgroup namespace has been added to systemd-nspawn. If
3409 supported by kernel, the container system started by systemd-nspawn
3410 will have its own view of the cgroup hierarchy. This new behaviour
3411 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
3413 * The new MemorySwapMax= option can be used to limit the maximum swap
3414 usage under the unified cgroup hierarchy.
3416 * Support for the CPU controller in the unified cgroup hierarchy has
3417 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
3418 options. This controller requires out-of-tree patches for the kernel
3419 and the support is provisional.
3421 * Mount and automount units may now be created transiently
3422 (i.e. dynamically at runtime via the bus API, instead of requiring
3423 unit files in the file system).
3425 * systemd-mount is a new tool which may mount file systems – much like
3426 mount(8), optionally pulling in additional dependencies through
3427 transient .mount and .automount units. For example, this tool
3428 automatically runs fsck on a backing block device before mounting,
3429 and allows the automount logic to be used dynamically from the
3430 command line for establishing mount points. This tool is particularly
3431 useful when dealing with removable media, as it will ensure fsck is
3432 run – if necessary – before the first access and that the file system
3433 is quickly unmounted after each access by utilizing the automount
3434 logic. This maximizes the chance that the file system on the
3435 removable media stays in a clean state, and if it isn't in a clean
3436 state is fixed automatically.
3438 * LazyUnmount=yes option for mount units has been added to expose the
3439 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
3442 * /efi will be used as the mount point of the EFI boot partition, if
3443 the directory is present, and the mount point was not configured
3444 through other means (e.g. fstab). If /efi directory does not exist,
3445 /boot will be used as before. This makes it easier to automatically
3446 mount the EFI partition on systems where /boot is used for something
3449 * When operating on GPT disk images for containers, systemd-nspawn will
3450 now mount the ESP to /boot or /efi according to the same rules as PID
3451 1 running on a host. This allows tools like "bootctl" to operate
3452 correctly within such containers, in order to make container images
3453 bootable on physical systems.
3455 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
3457 * Two new user session targets have been added to support running
3458 graphical sessions under the systemd --user instance:
3459 graphical-session.target and graphical-session-pre.target. See
3460 systemd.special(7) for a description of how those targets should be
3463 * The vconsole initialization code has been significantly reworked to
3464 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
3465 support unicode keymaps. Font and keymap configuration will now be
3466 copied to all allocated virtual consoles.
3468 * FreeBSD's bhyve virtualization is now detected.
3470 * Information recorded in the journal for core dumps now includes the
3471 contents of /proc/mountinfo and the command line of the process at
3472 the top of the process hierarchy (which is usually the init process
3475 * systemd-journal-gatewayd learned the --directory= option to serve
3476 files from the specified location.
3478 * journalctl --root=… can be used to peruse the journal in the
3479 /var/log/ directories inside of a container tree. This is similar to
3480 the existing --machine= option, but does not require the container to
3483 * The hardware database has been extended to support
3484 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
3487 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
3488 specify the click rate for mice which include a horizontal wheel with
3489 a click rate that is different than the one for the vertical wheel.
3491 * systemd-run gained a new --wait option that makes service execution
3492 synchronous. (Specifically, the command will not return until the
3493 specified service binary exited.)
3495 * systemctl gained a new --wait option that causes the start command to
3496 wait until the units being started have terminated again.
3498 * A new journal output mode "short-full" has been added which displays
3499 timestamps with abbreviated English day names and adds a timezone
3500 suffix. Those timestamps include more information than the default
3501 "short" output mode, and can be passed directly to journalctl's
3502 --since= and --until= options.
3504 * /etc/resolv.conf will be bind-mounted into containers started by
3505 systemd-nspawn, if possible, so any changes to resolv.conf contents
3506 are automatically propagated to the container.
3508 * The number of instances for socket-activated services originating
3509 from a single IP address can be limited with
3510 MaxConnectionsPerSource=, extending the existing setting of
3513 * systemd-networkd gained support for vcan ("Virtual CAN") interface
3516 * .netdev and .network configuration can now be extended through
3519 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
3520 Segmentation Offload, Generic Receive Offload, Large Receive Offload
3521 can be enabled and disabled using the new UDPSegmentationOffload=,
3522 TCPSegmentationOffload=, GenericSegmentationOffload=,
3523 GenericReceiveOffload=, LargeReceiveOffload= options in the
3524 [Link] section of .link files.
3526 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
3527 Port VLAN ID can be configured for bridge devices using the new STP=,
3528 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
3529 section of .netdev files.
3531 * The route table to which routes received over DHCP or RA should be
3532 added can be configured with the new RouteTable= option in the [DHCP]
3533 and [IPv6AcceptRA] sections of .network files.
3535 * The Address Resolution Protocol can be disabled on links managed by
3536 systemd-networkd using the ARP=no setting in the [Link] section of
3539 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
3540 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
3541 encode information about the result and exit codes of the current
3542 service runtime cycle.
3544 * systemd-sysctl will now configure kernel parameters in the order
3545 they occur in the configuration files. This matches what sysctl
3546 has been traditionally doing.
3548 * kernel-install "plugins" that are executed to perform various
3549 tasks after a new kernel is added and before an old one is removed
3550 can now return a special value to terminate the procedure and
3551 prevent any later plugins from running.
3553 * Journald's SplitMode=login setting has been deprecated. It has been
3554 removed from documentation, and its use is discouraged. In a future
3555 release it will be completely removed, and made equivalent to current
3556 default of SplitMode=uid.
3558 * Storage=both option setting in /etc/systemd/coredump.conf has been
3559 removed. With fast LZ4 compression storing the core dump twice is not
3562 * The --share-system systemd-nspawn option has been replaced with an
3563 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
3564 this functionality is discouraged. In addition the variables
3565 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
3566 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
3567 individual namespaces.
3569 * "machinectl list" now shows the IP address of running containers in
3570 the output, as well as OS release information.
3572 * "loginctl list" now shows the TTY of each session in the output.
3574 * sd-bus gained new API calls sd_bus_track_set_recursive(),
3575 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
3576 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
3577 tracking objects in a "recursive" mode, where a single client can be
3578 counted multiple times, if it takes multiple references.
3580 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
3581 sd_bus_get_exit_on_disconnect(). They may be used to make a
3582 process using sd-bus automatically exit if the bus connection is
3585 * Bus clients of the service manager may now "pin" loaded units into
3586 memory, by taking an explicit reference on them. This is useful to
3587 ensure the client can retrieve runtime data about the service even
3588 after the service completed execution. Taking such a reference is
3589 available only for privileged clients and should be helpful to watch
3590 running services in a race-free manner, and in particular collect
3591 information about exit statuses and results.
3593 * The nss-resolve module has been changed to strictly return UNAVAIL
3594 when communication via D-Bus with resolved failed, and NOTFOUND when
3595 a lookup completed but was negative. This means it is now possible to
3596 neatly configure fallbacks using nsswitch.conf result checking
3597 expressions. Taking benefit of this, the new recommended
3598 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
3600 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
3602 * A new setting CtrlAltDelBurstAction= has been added to
3603 /etc/systemd/system.conf which may be used to configure the precise
3604 behaviour if the user on the console presses Ctrl-Alt-Del more often
3605 than 7 times in 2s. Previously this would unconditionally result in
3606 an expedited, immediate reboot. With this new setting the precise
3607 operation may be configured in more detail, and also turned off
3610 * In .netdev files two new settings RemoteChecksumTx= and
3611 RemoteChecksumRx= are now understood that permit configuring the
3612 remote checksumming logic for VXLAN networks.
3614 * The service manager learnt a new "invocation ID" concept for invoked
3615 services. Each runtime cycle of a service will get a new invocation
3616 ID (a 128bit random UUID) assigned that identifies the current
3617 run of the service uniquely and globally. A new invocation ID
3618 is generated each time a service starts up. The journal will store
3619 the invocation ID of a service along with any logged messages, thus
3620 making the invocation ID useful for matching the online runtime of a
3621 service with the offline log data it generated in a safe way without
3622 relying on synchronized timestamps. In many ways this new service
3623 invocation ID concept is similar to the kernel's boot ID concept that
3624 uniquely and globally identifies the runtime of each boot. The
3625 invocation ID of a service is passed to the service itself via an
3626 environment variable ($INVOCATION_ID). A new bus call
3627 GetUnitByInvocationID() has been added that is similar to GetUnit()
3628 but instead of retrieving the bus path for a unit by its name
3629 retrieves it by its invocation ID. The returned path is valid only as
3630 long as the passed invocation ID is current.
3632 * systemd-resolved gained a new "DNSStubListener" setting in
3633 resolved.conf. It either takes a boolean value or the special values
3634 "udp" and "tcp", and configures whether to enable the stub DNS
3635 listener on 127.0.0.53:53.
3637 * IP addresses configured via networkd may now carry additional
3638 configuration settings supported by the kernel. New options include:
3639 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
3640 PrefixRoute=, AutoJoin=.
3642 * The PAM configuration fragment file for "user@.service" shipped with
3643 systemd (i.e. the --user instance of systemd) has been stripped to
3644 the minimum necessary to make the system boot. Previously, it
3645 contained Fedora-specific stanzas that did not apply to other
3646 distributions. It is expected that downstream distributions add
3647 additional configuration lines, matching their needs to this file,
3648 using it only as rough template of what systemd itself needs. Note
3649 that this reduced fragment does not even include an invocation of
3650 pam_limits which most distributions probably want to add, even though
3651 systemd itself does not need it. (There's also the new build time
3652 option --with-pamconfdir=no to disable installation of the PAM
3655 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
3656 capability is now also dropped from its set (in addition to
3657 CAP_SYS_MKNOD as before).
3659 * In service unit files it is now possible to connect a specific named
3660 file descriptor with stdin/stdout/stdout of an executed service. The
3661 name may be specified in matching .socket units using the
3662 FileDescriptorName= setting.
3664 * A number of journal settings may now be configured on the kernel
3665 command line. Specifically, the following options are now understood:
3666 systemd.journald.max_level_console=,
3667 systemd.journald.max_level_store=,
3668 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
3669 systemd.journald.max_level_wall=.
3671 * "systemctl is-enabled --full" will now show by which symlinks a unit
3672 file is enabled in the unit dependency tree.
3674 * Support for VeraCrypt encrypted partitions has been added to the
3675 "cryptsetup" logic and /etc/crypttab.
3677 * systemd-detect-virt gained support for a new --private-users switch
3678 that checks whether the invoking processes are running inside a user
3679 namespace. Similar, a new special value "private-users" for the
3680 existing ConditionVirtualization= setting has been added, permitting
3681 skipping of specific units in user namespace environments.
3683 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
3684 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
3685 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
3686 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
3687 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
3688 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
3689 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
3690 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
3691 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
3692 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
3693 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
3694 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
3695 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
3696 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
3697 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
3698 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
3699 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
3700 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
3701 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
3702 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
3703 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
3704 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
3705 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
3706 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
3707 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
3708 Jędrzejewski-Szmek, Zeal Jagannatha
3710 — Santa Fe, 2016-11-03
3714 * In service units the various ExecXYZ= settings have been extended
3715 with an additional special character as first argument of the
3716 assigned value: if the character '+' is used the specified command
3717 line it will be run with full privileges, regardless of User=,
3718 Group=, CapabilityBoundingSet= and similar options. The effect is
3719 similar to the existing PermissionsStartOnly= option, but allows
3720 configuration of this concept for each executed command line
3723 * Services may now alter the service watchdog timeout at runtime by
3724 sending a WATCHDOG_USEC= message via sd_notify().
3726 * MemoryLimit= and related unit settings now optionally take percentage
3727 specifications. The percentage is taken relative to the amount of
3728 physical memory in the system (or in case of containers, the assigned
3729 amount of memory). This allows scaling service resources neatly with
3730 the amount of RAM available on the system. Similarly, systemd-logind's
3731 RuntimeDirectorySize= option now also optionally takes percentage
3734 * In similar fashion TasksMax= takes percentage values now, too. The
3735 value is taken relative to the configured maximum number of processes
3736 on the system. The per-service task maximum has been changed to 15%
3737 using this functionality. (Effectively this is an increase of 512 →
3738 4915 for service units, given the kernel's default pid_max setting.)
3740 * Calendar time specifications in .timer units now understand a ".."
3741 syntax for time ranges. Example: "4..7:10" may now be used for
3742 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
3745 * The InaccessableDirectories=, ReadOnlyDirectories= and
3746 ReadWriteDirectories= unit file settings have been renamed to
3747 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
3748 applied to all kinds of file nodes, and not just directories, with
3749 the exception of symlinks. Specifically these settings may now be
3750 used on block and character device nodes, UNIX sockets and FIFOS as
3751 well as regular files. The old names of these settings remain
3752 available for compatibility.
3754 * systemd will now log about all service processes it kills forcibly
3755 (using SIGKILL) because they remained after the clean shutdown phase
3756 of the service completed. This should help identifying services that
3757 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
3758 systemd-logind's configuration a similar log message is generated for
3759 processes killed at the end of each session due to this setting.
3761 * systemd will now set the $JOURNAL_STREAM environment variable for all
3762 services whose stdout/stderr are connected to the Journal (which
3763 effectively means by default: all services). The variable contains
3764 the device and inode number of the file descriptor used for
3765 stdout/stderr. This may be used by invoked programs to detect whether
3766 their stdout/stderr is connected to the Journal, in which case they
3767 can switch over to direct Journal communication, thus being able to
3768 pass extended, structured metadata along with their log messages. As
3769 one example, this is now used by glib's logging primitives.
3771 * When using systemd's default tmp.mount unit for /tmp, the mount point
3772 will now be established with the "nosuid" and "nodev" options. This
3773 avoids privilege escalation attacks that put traps and exploits into
3774 /tmp. However, this might cause problems if you e. g. put container
3775 images or overlays into /tmp; if you need this, override tmp.mount's
3776 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
3779 * systemd now supports the "memory" cgroup controller also on
3782 * The systemd-cgtop tool now optionally takes a control group path as
3783 command line argument. If specified, the control group list shown is
3784 limited to subgroups of that group.
3786 * The SystemCallFilter= unit file setting gained support for
3787 pre-defined, named system call filter sets. For example
3788 SystemCallFilter=@clock is now an effective way to make all clock
3789 changing-related system calls unavailable to a service. A number of
3790 similar pre-defined groups are defined. Writing system call filters
3791 for system services is simplified substantially with this new
3792 concept. Accordingly, all of systemd's own, long-running services now
3793 enable system call filtering based on this, by default.
3795 * A new service setting MemoryDenyWriteExecute= has been added, taking
3796 a boolean value. If turned on, a service may no longer create memory
3797 mappings that are writable and executable at the same time. This
3798 enhances security for services where this is enabled as it becomes
3799 harder to dynamically write and then execute memory in exploited
3800 service processes. This option has been enabled for all of systemd's
3801 own long-running services.
3803 * A new RestrictRealtime= service setting has been added, taking a
3804 boolean argument. If set the service's processes may no longer
3805 acquire realtime scheduling. This improves security as realtime
3806 scheduling may otherwise be used to easily freeze the system.
3808 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
3809 value. This may be used for requesting that the system manager inside
3810 of the container reports start-up completion to nspawn which then
3811 propagates this notification further to the service manager
3812 supervising nspawn itself. A related option NotifyReady= in .nspawn
3813 files has been added too. This functionality allows ordering of the
3814 start-up of multiple containers using the usual systemd ordering
3817 * machinectl gained a new command "stop" that is an alias for
3820 * systemd-resolved gained support for contacting DNS servers on
3821 link-local IPv6 addresses.
3823 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
3824 its caches. A method call for requesting the same operation has been
3825 added to the bus API too, and is made available via "systemd-resolve
3828 * systemd-resolve gained a new --status switch. If passed a brief
3829 summary of the used DNS configuration with per-interface information
3832 * resolved.conf gained a new Cache= boolean option, defaulting to
3833 on. If turned off local DNS caching is disabled. This comes with a
3834 performance penalty in particular when DNSSEC is enabled. Note that
3835 resolved disables its internal caching implicitly anyway, when the
3836 configured DNS server is on a host-local IP address such as ::1 or
3837 127.0.0.1, thus automatically avoiding double local caching.
3839 * systemd-resolved now listens on the local IP address 127.0.0.53:53
3840 for DNS requests. This improves compatibility with local programs
3841 that do not use the libc NSS or systemd-resolved's bus APIs for name
3842 resolution. This minimal DNS service is only available to local
3843 programs and does not implement the full DNS protocol, but enough to
3844 cover local DNS clients. A new, static resolv.conf file, listing just
3845 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
3846 now recommended to make /etc/resolv.conf a symlink to this file in
3847 order to route all DNS lookups to systemd-resolved, regardless if
3848 done via NSS, the bus API or raw DNS packets. Note that this local
3849 DNS service is not as fully featured as the libc NSS or
3850 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
3851 used to deliver link-local address information (as this implies
3852 sending a local interface index along), LLMNR/mDNS support via this
3853 interface is severely restricted. It is thus strongly recommended for
3854 all applications to use the libc NSS API or native systemd-resolved
3857 * systemd-networkd's bridge support learned a new setting
3858 VLANFiltering= for controlling VLAN filtering. Moreover a new section
3859 in .network files has been added for configuring VLAN bridging in
3860 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
3862 * systemd-networkd's IPv6 Router Advertisement code now makes use of
3863 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
3864 now be acquired without relying on DHCPv6. Two new options
3865 UseDomains= and UseDNS= have been added to configure this behaviour.
3867 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
3868 renamed IPv6AcceptRA=, without altering its behaviour. The old
3869 setting name remains available for compatibility reasons.
3871 * The systemd-networkd VTI/VTI6 tunneling support gained new options
3872 Key=, InputKey= and OutputKey=.
3874 * systemd-networkd gained support for VRF ("Virtual Routing Function")
3875 interface configuration.
3877 * "systemctl edit" may now be used to create new unit files by
3878 specifying the --force switch.
3880 * sd-event gained a new function sd_event_get_iteration() for
3881 requesting the current iteration counter of the event loop. It starts
3882 at zero and is increased by one with each event loop iteration.
3884 * A new rpm macro %systemd_ordering is provided by the macros.systemd
3885 file. It can be used in lieu of %systemd_requires in packages which
3886 don't use any systemd functionality and are intended to be installed
3887 in minimal containers without systemd present. This macro provides
3888 ordering dependencies to ensure that if the package is installed in
3889 the same rpm transaction as systemd, systemd will be installed before
3890 the scriptlets for the package are executed, allowing unit presets
3893 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
3894 been added to simplify packaging of generators.
3896 * The os-release file gained VERSION_CODENAME field for the
3897 distribution nickname (e.g. VERSION_CODENAME=woody).
3899 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
3900 can be set to disable parsing of metadata and the creation
3901 of persistent symlinks for that device.
3903 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
3904 to make them available to logged-in users has been reverted.
3906 * Much of the common code of the various systemd components is now
3907 built into an internal shared library libsystemd-shared-231.so
3908 (incorporating the systemd version number in the name, to be updated
3909 with future releases) that the components link to. This should
3910 decrease systemd footprint both in memory during runtime and on
3911 disk. Note that the shared library is not for public use, and is
3912 neither API nor ABI stable, but is likely to change with every new
3913 released update. Packagers need to make sure that binaries
3914 linking to libsystemd-shared.so are updated in step with the
3917 * Configuration for "mkosi" is now part of the systemd
3918 repository. mkosi is a tool to easily build legacy-free OS images,
3919 and is available on github: https://github.com/systemd/mkosi. If
3920 "mkosi" is invoked in the build tree a new raw OS image is generated
3921 incorporating the systemd sources currently being worked on and a
3922 clean, fresh distribution installation. The generated OS image may be
3923 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
3924 UEFI PC. This functionality is particularly useful to easily test
3925 local changes made to systemd in a pristine, defined environment. See
3926 doc/HACKING for details.
3928 * configure learned the --with-support-url= option to specify the
3929 distribution's bugtracker.
3931 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
3932 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
3933 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
3934 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
3935 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
3936 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
3937 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
3938 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
3939 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
3940 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
3941 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
3942 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
3943 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
3944 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
3945 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
3946 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
3947 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
3948 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
3949 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
3951 — Berlin, 2016-07-25
3955 * DNSSEC is now turned on by default in systemd-resolved (in
3956 "allow-downgrade" mode), but may be turned off during compile time by
3957 passing "--with-default-dnssec=no" to "configure" (and of course,
3958 during runtime with DNSSEC= in resolved.conf). We recommend
3959 downstreams to leave this on at least during development cycles and
3960 report any issues with the DNSSEC logic upstream. We are very
3961 interested in collecting feedback about the DNSSEC validator and its
3962 limitations in the wild. Note however, that DNSSEC support is
3963 probably nothing downstreams should turn on in stable distros just
3964 yet, as it might create incompatibilities with a few DNS servers and
3965 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
3966 automatically whenever we detect such incompatible setups, but there
3967 might be systems we do not cover yet. Hence: please help us testing
3968 the DNSSEC code, leave this on where you can, report back, but then
3969 again don't consider turning this on in your stable, LTS or
3970 production release just yet. (Note that you have to enable
3971 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
3972 and its DNSSEC mode for host name resolution from local
3975 * systemd-resolve conveniently resolves DANE records with the --tlsa
3976 option and OPENPGPKEY records with the --openpgp option. It also
3977 supports dumping raw DNS record data via the new --raw= switch.
3979 * systemd-logind will now by default terminate user processes that are
3980 part of the user session scope unit (session-XX.scope) when the user
3981 logs out. This behavior is controlled by the KillUserProcesses=
3982 setting in logind.conf, and the previous default of "no" is now
3983 changed to "yes". This means that user sessions will be properly
3984 cleaned up after, but additional steps are necessary to allow
3985 intentionally long-running processes to survive logout.
3987 While the user is logged in at least once, user@.service is running,
3988 and any service that should survive the end of any individual login
3989 session can be started at a user service or scope using systemd-run.
3990 systemd-run(1) man page has been extended with an example which shows
3991 how to run screen in a scope unit underneath user@.service. The same
3992 command works for tmux.
3994 After the user logs out of all sessions, user@.service will be
3995 terminated too, by default, unless the user has "lingering" enabled.
3996 To effectively allow users to run long-term tasks even if they are
3997 logged out, lingering must be enabled for them. See loginctl(1) for
3998 details. The default polkit policy was modified to allow users to
3999 set lingering for themselves without authentication.
4001 Previous defaults can be restored at compile time by the
4002 --without-kill-user-processes option to "configure".
4004 * systemd-logind gained new configuration settings SessionsMax= and
4005 InhibitorsMax=, both with a default of 8192. It will not register new
4006 user sessions or inhibitors above this limit.
4008 * systemd-logind will now reload configuration on SIGHUP.
4010 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
4011 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
4012 enable. Also, support for the "io" cgroup controller in the unified
4013 hierarchy has been added, so that the "memory", "pids" and "io" are
4014 now the controllers that are supported on the unified hierarchy.
4016 WARNING: it is not possible to use previous systemd versions with
4017 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
4018 is necessary to also update systemd in the initramfs if using the
4019 unified hierarchy. An updated SELinux policy is also required.
4021 * LLDP support has been extended, and both passive (receive-only) and
4022 active (sender) modes are supported. Passive mode ("routers-only") is
4023 enabled by default in systemd-networkd. Active LLDP mode is enabled
4024 by default for containers on the internal network. The "networkctl
4025 lldp" command may be used to list information gathered. "networkctl
4026 status" will also show basic LLDP information on connected peers now.
4028 * The IAID and DUID unique identifier sent in DHCP requests may now be
4029 configured for the system and each .network file managed by
4030 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
4032 * systemd-networkd gained support for configuring proxy ARP support for
4033 each interface, via the ProxyArp= setting in .network files. It also
4034 gained support for configuring the multicast querier feature of
4035 bridge devices, via the new MulticastQuerier= setting in .netdev
4036 files. Similarly, snooping on the IGMP traffic can be controlled
4037 via the new setting MulticastSnooping=.
4039 A new setting PreferredLifetime= has been added for addresses
4040 configured in .network file to configure the lifetime intended for an
4043 The systemd-networkd DHCP server gained the option EmitRouter=, which
4044 defaults to yes, to configure whether the DHCP Option 3 (Router)
4047 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
4048 systemd-socket-activate and installed into /usr/bin. It is now fully
4051 * systemd-journald now uses separate threads to flush changes to disk
4052 when closing journal files, thus reducing impact of slow disk I/O on
4053 logging performance.
4055 * The sd-journal API gained two new calls
4056 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
4057 can be used to open journal files using file descriptors instead of
4058 file or directory paths. sd_journal_open_container() has been
4059 deprecated, sd_journal_open_directory_fd() should be used instead
4060 with the flag SD_JOURNAL_OS_ROOT.
4062 * journalctl learned a new output mode "-o short-unix" that outputs log
4063 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
4064 UTC). It also gained support for a new --no-hostname setting to
4065 suppress the hostname column in the family of "short" output modes.
4067 * systemd-ask-password now optionally skips printing of the password to
4068 stdout with --no-output which can be useful in scripts.
4070 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
4071 (devices tagged with ID_MAKER_TOOL) are now tagged with
4072 "uaccess" and are available to logged in users.
4074 * The DeviceAllow= unit setting now supports specifiers (with "%").
4076 * "systemctl show" gained a new --value switch, which allows print a
4077 only the contents of a specific unit property, without also printing
4078 the property's name. Similar support was added to "show*" verbs
4079 of loginctl and machinectl that output "key=value" lists.
4081 * A new unit type "generated" was added for files dynamically generated
4082 by generator tools. Similarly, a new unit type "transient" is used
4083 for unit files created using the runtime API. "systemctl enable" will
4084 refuse to operate on such files.
4086 * A new command "systemctl revert" has been added that may be used to
4087 revert to the vendor version of a unit file, in case local changes
4088 have been made by adding drop-ins or overriding the unit file.
4090 * "machinectl clean" gained a new verb to automatically remove all or
4091 just hidden container images.
4093 * systemd-tmpfiles gained support for a new line type "e" for emptying
4094 directories, if they exist, without creating them if they don't.
4096 * systemd-nspawn gained support for automatically patching the UID/GIDs
4097 of the owners and the ACLs of all files and directories in a
4098 container tree to match the UID/GID user namespacing range selected
4099 for the container invocation. This mode is enabled via the new
4100 --private-users-chown switch. It also gained support for
4101 automatically choosing a free, previously unused UID/GID range when
4102 starting a container, via the new --private-users=pick setting (which
4103 implies --private-users-chown). Together, these options for the first
4104 time make user namespacing for nspawn containers fully automatic and
4105 thus deployable. The systemd-nspawn@.service template unit file has
4106 been changed to use this functionality by default.
4108 * systemd-nspawn gained a new --network-zone= switch, that allows
4109 creating ad-hoc virtual Ethernet links between multiple containers,
4110 that only exist as long as at least one container referencing them is
4111 running. This allows easy connecting of multiple containers with a
4112 common link that implements an Ethernet broadcast domain. Each of
4113 these network "zones" may be named relatively freely by the user, and
4114 may be referenced by any number of containers, but each container may
4115 only reference one of these "zones". On the lower level, this is
4116 implemented by an automatically managed bridge network interface for
4117 each zone, that is created when the first container referencing its
4118 zone is created and removed when the last one referencing its zone
4121 * The default start timeout may now be configured on the kernel command
4122 line via systemd.default_timeout_start_sec=. It was already
4123 configurable via the DefaultTimeoutStartSec= option in
4124 /etc/systemd/system.conf.
4126 * Socket units gained a new TriggerLimitIntervalSec= and
4127 TriggerLimitBurst= setting to configure a limit on the activation
4128 rate of the socket unit.
4130 * The LimitNICE= setting now optionally takes normal UNIX nice values
4131 in addition to the raw integer limit value. If the specified
4132 parameter is prefixed with "+" or "-" and is in the range -20..19 the
4133 value is understood as UNIX nice value. If not prefixed like this it
4134 is understood as raw RLIMIT_NICE limit.
4136 * Note that the effect of the PrivateDevices= unit file setting changed
4137 slightly with this release: the per-device /dev file system will be
4138 mounted read-only from this version on, and will have "noexec"
4139 set. This (minor) change of behavior might cause some (exceptional)
4140 legacy software to break, when PrivateDevices=yes is set for its
4141 service. Please leave PrivateDevices= off if you run into problems
4144 * systemd-bootchart has been split out to a separate repository:
4145 https://github.com/systemd/systemd-bootchart
4147 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
4148 merged into the kernel in its current form.
4150 * The compatibility libraries libsystemd-daemon.so,
4151 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
4152 which have been deprecated since systemd-209 have been removed along
4153 with the corresponding pkg-config files. All symbols provided by
4154 those libraries are provided by libsystemd.so.
4156 * The Capabilities= unit file setting has been removed (it is ignored
4157 for backwards compatibility). AmbientCapabilities= and
4158 CapabilityBoundingSet= should be used instead.
4160 * A new special target has been added, initrd-root-device.target,
4161 which creates a synchronization point for dependencies of the root
4162 device in early userspace. Initramfs builders must ensure that this
4163 target is now included in early userspace.
4165 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
4166 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
4167 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
4168 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
4169 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
4170 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
4171 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
4172 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
4173 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
4174 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
4175 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
4176 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
4177 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
4178 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
4179 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
4180 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
4181 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
4182 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
4183 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
4184 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
4185 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
4186 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
4187 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
4188 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
4189 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
4192 — Fairfax, 2016-05-21
4196 * The systemd-resolved DNS resolver service has gained a substantial
4197 set of new features, most prominently it may now act as a DNSSEC
4198 validating stub resolver. DNSSEC mode is currently turned off by
4199 default, but is expected to be turned on by default in one of the
4200 next releases. For now, we invite everybody to test the DNSSEC logic
4201 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
4202 service also gained a full set of D-Bus interfaces, including calls
4203 to configure DNS and DNSSEC settings per link (for use by external
4204 network management software). systemd-resolved and systemd-networkd
4205 now distinguish between "search" and "routing" domains. The former
4206 are used to qualify single-label names, the latter are used purely
4207 for routing lookups within certain domains to specific links.
4208 resolved now also synthesizes RRs for all entries from /etc/hosts.
4210 * The systemd-resolve tool (which is a client utility for
4211 systemd-resolved) has been improved considerably and is now fully
4212 supported and documented. Hence it has moved from /usr/lib/systemd to
4215 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
4218 * The coredump collection logic has been reworked: when a coredump is
4219 collected it is now written to disk, compressed and processed
4220 (including stacktrace extraction) from a new instantiated service
4221 systemd-coredump@.service, instead of directly from the
4222 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
4223 processing large coredumps can take up a substantial amount of
4224 resources and time, and this previously happened entirely outside of
4225 systemd's service supervision. With the new logic the core_pattern
4226 hook only does minimal metadata collection before passing off control
4227 to the new instantiated service, which is configured with a time
4228 limit, a nice level and other settings to minimize negative impact on
4229 the rest of the system. Also note that the new logic will honour the
4230 RLIMIT_CORE setting of the crashed process, which now allows users
4231 and processes to turn off coredumping for their processes by setting
4234 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
4235 and all forked processes by default. Previously, PID 1 would leave
4236 the setting at "0" for all processes, as set by the kernel. Note that
4237 the resource limit traditionally has no effect on the generated
4238 coredumps on the system if the /proc/sys/kernel/core_pattern hook
4239 logic is used. Since the limit is now honoured (see above) its
4240 default has been changed so that the coredumping logic is enabled by
4241 default for all processes, while allowing specific opt-out.
4243 * When the stacktrace is extracted from processes of system users, this
4244 is now done as "systemd-coredump" user, in order to sandbox this
4245 potentially security sensitive parsing operation. (Note that when
4246 processing coredumps of normal users this is done under the user ID
4247 of process that crashed, as before.) Packagers should take notice
4248 that it is now necessary to create the "systemd-coredump" system user
4249 and group at package installation time.
4251 * The systemd-activate socket activation testing tool gained support
4252 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
4253 and --seqpacket switches. It also has been extended to support both
4254 new-style and inetd-style file descriptor passing. Use the new
4255 --inetd switch to request inetd-style file descriptor passing.
4257 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
4258 variable, which takes a boolean value. If set to false, ANSI color
4259 output is disabled in the tools even when run on a terminal that
4262 * The VXLAN support in networkd now supports two new settings
4263 DestinationPort= and PortRange=.
4265 * A new systemd.machine_id= kernel command line switch has been added,
4266 that may be used to set the machine ID in /etc/machine-id if it is
4267 not initialized yet. This command line option has no effect if the
4268 file is already initialized.
4270 * systemd-nspawn gained a new --as-pid2 switch that invokes any
4271 specified command line as PID 2 rather than PID 1 in the
4272 container. In this mode PID 1 is a minimal stub init process that
4273 implements the special POSIX and Linux semantics of PID 1 regarding
4274 signal and child process management. Note that this stub init process
4275 is implemented in nspawn itself and requires no support from the
4276 container image. This new logic is useful to support running
4277 arbitrary commands in the container, as normal processes are
4278 generally not prepared to run as PID 1.
4280 * systemd-nspawn gained a new --chdir= switch for setting the current
4281 working directory for the process started in the container.
4283 * "journalctl /dev/sda" will now output all kernel log messages for
4284 specified device from the current boot, in addition to all devices
4285 that are parents of it. This should make log output about devices
4286 pretty useful, as long as kernel drivers attach enough metadata to
4287 the log messages. (The usual SATA drivers do.)
4289 * The sd-journal API gained two new calls
4290 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
4291 that report whether log data from /run or /var has been found.
4293 * journalctl gained a new switch "--fields" that prints all journal
4294 record field names currently in use in the journal. This is backed
4295 by two new sd-journal API calls sd_journal_enumerate_fields() and
4296 sd_journal_restart_fields().
4298 * Most configurable timeouts in systemd now expect an argument of
4299 "infinity" to turn them off, instead of "0" as before. The semantics
4300 from now on is that a timeout of "0" means "now", and "infinity"
4301 means "never". To maintain backwards compatibility, "0" continues to
4302 turn off previously existing timeout settings.
4304 * "systemctl reload-or-try-restart" has been renamed to "systemctl
4305 try-reload-or-restart" to clarify what it actually does: the "try"
4306 logic applies to both reloading and restarting, not just restarting.
4307 The old name continues to be accepted for compatibility.
4309 * On boot-up, when PID 1 detects that the system clock is behind the
4310 release date of the systemd version in use, the clock is now set
4311 to the latter. Previously, this was already done in timesyncd, in order
4312 to avoid running with clocks set to the various clock epochs such as
4313 1902, 1938 or 1970. With this change the logic is now done in PID 1
4314 in addition to timesyncd during early boot-up, so that it is enforced
4315 before the first process is spawned by systemd. Note that the logic
4316 in timesyncd remains, as it is more comprehensive and ensures
4317 clock monotonicity by maintaining a persistent timestamp file in
4318 /var. Since /var is generally not available in earliest boot or the
4319 initrd, this part of the logic remains in timesyncd, and is not done
4322 * Support for tweaking details in net_cls.class_id through the
4323 NetClass= configuration directive has been removed, as the kernel
4324 people have decided to deprecate that controller in cgroup v2.
4325 Userspace tools such as nftables are moving over to setting rules
4326 that are specific to the full cgroup path of a task, which obsoletes
4327 these controllers anyway. The NetClass= directive is kept around for
4328 legacy compatibility reasons. For a more in-depth description of the
4329 kernel change, please refer to the respective upstream commit:
4331 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
4333 * A new service setting RuntimeMaxSec= has been added that may be used
4334 to specify a maximum runtime for a service. If the timeout is hit, the
4335 service is terminated and put into a failure state.
4337 * A new service setting AmbientCapabilities= has been added. It allows
4338 configuration of additional Linux process capabilities that are
4339 passed to the activated processes. This is only available on very
4342 * The process resource limit settings in service units may now be used
4343 to configure hard and soft limits individually.
4345 * The various libsystemd APIs such as sd-bus or sd-event now publicly
4346 expose support for gcc's __attribute__((cleanup())) C extension.
4347 Specifically, for many object destructor functions alternative
4348 versions have been added that have names suffixed with "p" and take a
4349 pointer to a pointer to the object to destroy, instead of just a
4350 pointer to the object itself. This is useful because these destructor
4351 functions may be used directly as parameters to the cleanup
4352 construct. Internally, systemd has been a heavy user of this GCC
4353 extension for a long time, and with this change similar support is
4354 now available to consumers of the library outside of systemd. Note
4355 that by using this extension in your sources compatibility with old
4356 and strictly ANSI compatible C compilers is lost. However, all gcc or
4357 LLVM versions of recent years support this extension.
4359 * Timer units gained support for a new setting RandomizedDelaySec= that
4360 allows configuring some additional randomized delay to the configured
4361 time. This is useful to spread out timer events to avoid load peaks in
4362 clusters or larger setups.
4364 * Calendar time specifications now support sub-second accuracy.
4366 * Socket units now support listening on SCTP and UDP-lite protocol
4369 * The sd-event API now comes with a full set of man pages.
4371 * Older versions of systemd contained experimental support for
4372 compressing journal files and coredumps with the LZ4 compressor that
4373 was not compatible with the lz4 binary (due to API limitations of the
4374 lz4 library). This support has been removed; only support for files
4375 compatible with the lz4 binary remains. This LZ4 logic is now
4376 officially supported and no longer considered experimental.
4378 * The dkr image import logic has been removed again from importd. dkr's
4379 micro-services focus doesn't fit into the machine image focus of
4380 importd, and quickly got out of date with the upstream dkr API.
4382 * Creation of the /run/lock/lockdev/ directory was dropped from
4383 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
4384 been available for many years. If you still need this, you need to
4385 create your own tmpfiles.d config file with:
4387 d /run/lock/lockdev 0775 root lock -
4389 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
4390 and RebootArgument= have been moved from the [Service] section of
4391 unit files to [Unit], and they are now supported on all unit types,
4392 not just service units. Of course, systemd will continue to
4393 understand these settings also at the old location, in order to
4394 maintain compatibility.
4396 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
4397 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
4398 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
4399 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
4400 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
4401 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
4402 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
4403 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
4404 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
4405 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
4406 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
4407 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
4408 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
4409 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
4410 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
4411 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
4412 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
4413 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
4414 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
4416 — Berlin, 2016-02-11
4420 * A number of properties previously only settable in unit
4421 files are now also available as properties to set when
4422 creating transient units programmatically via the bus, as it
4423 is exposed with systemd-run's --property=
4424 setting. Specifically, these are: SyslogIdentifier=,
4425 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
4426 EnvironmentFile=, ReadWriteDirectories=,
4427 ReadOnlyDirectories=, InaccessibleDirectories=,
4428 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
4430 * When creating transient services via the bus API it is now
4431 possible to pass in a set of file descriptors to use as
4432 STDIN/STDOUT/STDERR for the invoked process.
4434 * Slice units may now be created transiently via the bus APIs,
4435 similar to the way service and scope units may already be
4436 created transiently.
4438 * Wherever systemd expects a calendar timestamp specification
4439 (like in journalctl's --since= and --until= switches) UTC
4440 timestamps are now supported. Timestamps suffixed with "UTC"
4441 are now considered to be in Universal Time Coordinated
4442 instead of the local timezone. Also, timestamps may now
4443 optionally be specified with sub-second accuracy. Both of
4444 these additions also apply to recurring calendar event
4445 specification, such as OnCalendar= in timer units.
4447 * journalctl gained a new "--sync" switch that asks the
4448 journal daemon to write all so far unwritten log messages to
4449 disk and sync the files, before returning.
4451 * systemd-tmpfiles learned two new line types "q" and "Q" that
4452 operate like "v", but also set up a basic btrfs quota
4453 hierarchy when used on a btrfs file system with quota
4456 * tmpfiles' "v", "q" and "Q" will now create a plain directory
4457 instead of a subvolume (even on a btrfs file system) if the
4458 root directory is a plain directory, and not a
4459 subvolume. This should simplify things with certain chroot()
4460 environments which are not aware of the concept of btrfs
4463 * systemd-detect-virt gained a new --chroot switch to detect
4464 whether execution takes place in a chroot() environment.
4466 * CPUAffinity= now takes CPU index ranges in addition to
4469 * The various memory-related resource limit settings (such as
4470 LimitAS=) now understand the usual K, M, G, ... suffixes to
4471 the base of 1024 (IEC). Similar, the time-related resource
4472 limit settings understand the usual min, h, day, ...
4475 * There's a new system.conf setting DefaultTasksMax= to
4476 control the default TasksMax= setting for services and
4477 scopes running on the system. (TasksMax= is the primary
4478 setting that exposes the "pids" cgroup controller on systemd
4479 and was introduced in the previous systemd release.) The
4480 setting now defaults to 512, which means services that are
4481 not explicitly configured otherwise will only be able to
4482 create 512 processes or threads at maximum, from this
4483 version on. Note that this means that thread- or
4484 process-heavy services might need to be reconfigured to set
4485 TasksMax= to a higher value. It is sufficient to set
4486 TasksMax= in these specific unit files to a higher value, or
4487 even "infinity". Similar, there's now a logind.conf setting
4488 UserTasksMax= that defaults to 4096 and limits the total
4489 number of processes or tasks each user may own
4490 concurrently. nspawn containers also have the TasksMax=
4491 value set by default now, to 8192. Note that all of this
4492 only has an effect if the "pids" cgroup controller is
4493 enabled in the kernel. The general benefit of these changes
4494 should be a more robust and safer system, that provides a
4495 certain amount of per-service fork() bomb protection.
4497 * systemd-nspawn gained the new --network-veth-extra= switch
4498 to define additional and arbitrarily-named virtual Ethernet
4499 links between the host and the container.
4501 * A new service execution setting PassEnvironment= has been
4502 added that allows importing select environment variables
4503 from PID1's environment block into the environment block of
4506 * Timer units gained support for a new RemainAfterElapse=
4507 setting which takes a boolean argument. It defaults to on,
4508 exposing behaviour unchanged to previous releases. If set to
4509 off, timer units are unloaded after they elapsed if they
4510 cannot elapse again. This is particularly useful for
4511 transient timer units, which shall not stay around longer
4512 than until they first elapse.
4514 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
4515 default now (the kernel default is 16). This is beneficial
4516 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
4517 allows substantially larger numbers of queued
4518 datagrams. This should increase the capability of systemd to
4519 parallelize boot-up, as logging and sd_notify() are unlikely
4520 to stall execution anymore. If you need to change the value
4521 from the new defaults, use the usual sysctl.d/ snippets.
4523 * The compression framing format used by the journal or
4524 coredump processing has changed to be in line with what the
4525 official LZ4 tools generate. LZ4 compression support in
4526 systemd was considered unsupported previously, as the format
4527 was not compatible with the normal tools. With this release
4528 this has changed now, and it is hence safe for downstream
4529 distributions to turn it on. While not compressing as well
4530 as the XZ, LZ4 is substantially faster, which makes
4531 it a good default choice for the compression logic in the
4532 journal and in coredump handling.
4534 * Any reference to /etc/mtab has been dropped from
4535 systemd. The file has been obsolete since a while, but
4536 systemd refused to work on systems where it was incorrectly
4537 set up (it should be a symlink or non-existent). Please make
4538 sure to update to util-linux 2.27.1 or newer in conjunction
4539 with this systemd release, which also drops any reference to
4540 /etc/mtab. If you maintain a distribution make sure that no
4541 software you package still references it, as this is a
4542 likely source of bugs. There's also a glibc bug pending,
4543 asking for removal of any reference to this obsolete file:
4545 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
4547 Note that only util-linux versions built with
4548 --enable-libmount-force-mountinfo are supported.
4550 * Support for the ".snapshot" unit type has been removed. This
4551 feature turned out to be little useful and little used, and
4552 has now been removed from the core and from systemctl.
4554 * The dependency types RequiresOverridable= and
4555 RequisiteOverridable= have been removed from systemd. They
4556 have been used only very sparingly to our knowledge and
4557 other options that provide a similar effect (such as
4558 systemctl --mode=ignore-dependencies) are much more useful
4559 and commonly used. Moreover, they were only half-way
4560 implemented as the option to control behaviour regarding
4561 these dependencies was never added to systemctl. By removing
4562 these dependency types the execution engine becomes a bit
4563 simpler. Unit files that use these dependencies should be
4564 changed to use the non-Overridable dependency types
4565 instead. In fact, when parsing unit files with these
4566 options, that's what systemd will automatically convert them
4567 too, but it will also warn, asking users to fix the unit
4568 files accordingly. Removal of these dependency types should
4569 only affect a negligible number of unit files in the wild.
4571 * Behaviour of networkd's IPForward= option changed
4572 (again). It will no longer maintain a per-interface setting,
4573 but propagate one way from interfaces where this is enabled
4574 to the global kernel setting. The global setting will be
4575 enabled when requested by a network that is set up, but
4576 never be disabled again. This change was made to make sure
4577 IPv4 and IPv6 behaviour regarding packet forwarding is
4578 similar (as the Linux IPv6 stack does not support
4579 per-interface control of this setting) and to minimize
4582 * In unit files the behaviour of %u, %U, %h, %s has
4583 changed. These specifiers will now unconditionally resolve
4584 to the various user database fields of the user that the
4585 systemd instance is running as, instead of the user
4586 configured in the specific unit via User=. Note that this
4587 effectively doesn't change much, as resolving of these
4588 specifiers was already turned off in the --system instance
4589 of systemd, as we cannot do NSS lookups from PID 1. In the
4590 --user instance of systemd these specifiers where correctly
4591 resolved, but hardly made any sense, since the user instance
4592 lacks privileges to do user switches anyway, and User= is
4593 hence useless. Moreover, even in the --user instance of
4594 systemd behaviour was awkward as it would only take settings
4595 from User= assignment placed before the specifier into
4596 account. In order to unify and simplify the logic around
4597 this the specifiers will now always resolve to the
4598 credentials of the user invoking the manager (which in case
4599 of PID 1 is the root user).
4601 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
4602 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
4603 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
4604 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
4605 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
4606 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
4607 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
4608 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
4609 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
4610 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
4611 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
4612 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
4613 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
4614 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
4617 — Berlin, 2015-11-18
4621 * systemd now depends on util-linux v2.27. More specifically,
4622 the newly added mount monitor feature in libmount now
4623 replaces systemd's former own implementation.
4625 * libmount mandates /etc/mtab not to be regular file, and
4626 systemd now enforces this condition at early boot.
4627 /etc/mtab has been deprecated and warned about for a very
4628 long time, so systems running systemd should already have
4629 stopped having this file around as anything else than a
4630 symlink to /proc/self/mounts.
4632 * Support for the "pids" cgroup controller has been added. It
4633 allows accounting the number of tasks in a cgroup and
4634 enforcing limits on it. This adds two new setting
4635 TasksAccounting= and TasksMax= to each unit, as well as a
4636 global option DefaultTasksAccounting=.
4638 * Support for the "net_cls" cgroup controller has been added.
4639 It allows assigning a net class ID to each task in the
4640 cgroup, which can then be used in firewall rules and traffic
4641 shaping configurations. Note that the kernel netfilter net
4642 class code does not currently work reliably for ingress
4643 packets on unestablished sockets.
4645 This adds a new config directive called NetClass= to CGroup
4646 enabled units. Allowed values are positive numbers for fixed
4647 assignments and "auto" for picking a free value
4650 * 'systemctl is-system-running' now returns 'offline' if the
4651 system is not booted with systemd. This command can now be
4652 used as a substitute for 'systemd-notify --booted'.
4654 * Watchdog timeouts have been increased to 3 minutes for all
4655 in-tree service files. Apparently, disk IO issues are more
4656 frequent than we hoped, and user reported >1 minute waiting
4659 * 'machine-id-commit' functionality has been merged into
4660 'machine-id-setup --commit'. The separate binary has been
4663 * The WorkingDirectory= directive in unit files may now be set
4664 to the special value '~'. In this case, the working
4665 directory is set to the home directory of the user
4666 configured in User=.
4668 * "machinectl shell" will now open the shell in the home
4669 directory of the selected user by default.
4671 * The CrashChVT= configuration file setting is renamed to
4672 CrashChangeVT=, following our usual logic of not
4673 abbreviating unnecessarily. The old directive is still
4674 supported for compat reasons. Also, this directive now takes
4675 an integer value between 1 and 63, or a boolean value. The
4676 formerly supported '-1' value for disabling stays around for
4679 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
4680 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
4681 RootDirectory= properties can now be set for transient
4684 * The systemd-analyze tool gained a new "set-log-target" verb
4685 to change the logging target the system manager logs to
4686 dynamically during runtime. This is similar to how
4687 "systemd-analyze set-log-level" already changes the log
4690 * In nspawn /sys is now mounted as tmpfs, with only a selected
4691 set of subdirectories mounted in from the real sysfs. This
4692 enhances security slightly, and is useful for ensuring user
4693 namespaces work correctly.
4695 * Support for USB FunctionFS activation has been added. This
4696 allows implementation of USB gadget services that are
4697 activated as soon as they are requested, so that they don't
4698 have to run continuously, similar to classic socket
4701 * The "systemctl exit" command now optionally takes an
4702 additional parameter that sets the exit code to return from
4703 the systemd manager when exiting. This is only relevant when
4704 running the systemd user instance, or when running the
4705 system instance in a container.
4707 * sd-bus gained the new API calls sd_bus_path_encode_many()
4708 and sd_bus_path_decode_many() that allow easy encoding and
4709 decoding of multiple identifier strings inside a D-Bus
4710 object path. Another new call sd_bus_default_flush_close()
4711 has been added to flush and close per-thread default
4714 * systemd-cgtop gained support for a -M/--machine= switch to
4715 show the control groups within a certain container only.
4717 * "systemctl kill" gained support for an optional --fail
4718 switch. If specified the requested operation will fail of no
4719 processes have been killed, because the unit had no
4720 processes attached, or similar.
4722 * A new systemd.crash_reboot=1 kernel command line option has
4723 been added that triggers a reboot after crashing. This can
4724 also be set through CrashReboot= in systemd.conf.
4726 * The RuntimeDirectory= setting now understands unit
4727 specifiers like %i or %f.
4729 * A new (still internal) library API sd-ipv4acd has been added,
4730 that implements address conflict detection for IPv4. It's
4731 based on code from sd-ipv4ll, and will be useful for
4732 detecting DHCP address conflicts.
4734 * File descriptors passed during socket activation may now be
4735 named. A new API sd_listen_fds_with_names() is added to
4736 access the names. The default names may be overridden,
4737 either in the .socket file using the FileDescriptorName=
4738 parameter, or by passing FDNAME= when storing the file
4739 descriptors using sd_notify().
4741 * systemd-networkd gained support for:
4743 - Setting the IPv6 Router Advertisement settings via
4744 IPv6AcceptRouterAdvertisements= in .network files.
4746 - Configuring the HelloTimeSec=, MaxAgeSec= and
4747 ForwardDelaySec= bridge parameters in .netdev files.
4749 - Configuring PreferredSource= for static routes in
4752 * The "ask-password" framework used to query for LUKS harddisk
4753 passwords or SSL passwords during boot gained support for
4754 caching passwords in the kernel keyring, if it is
4755 available. This makes sure that the user only has to type in
4756 a passphrase once if there are multiple objects to unlock
4757 with the same one. Previously, such password caching was
4758 available only when Plymouth was used; this moves the
4759 caching logic into the systemd codebase itself. The
4760 "systemd-ask-password" utility gained a new --keyname=
4761 switch to control which kernel keyring key to use for
4762 caching a password in. This functionality is also useful for
4763 enabling display managers such as gdm to automatically
4764 unlock the user's GNOME keyring if its passphrase, the
4765 user's password and the harddisk password are the same, if
4766 gdm-autologin is used.
4768 * When downloading tar or raw images using "machinectl
4769 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
4770 file is now also downloaded, if it is available and stored
4771 next to the image file.
4773 * Units of type ".socket" gained a new boolean setting
4774 Writable= which is only useful in conjunction with
4775 ListenSpecial=. If true, enables opening the specified
4776 special file in O_RDWR mode rather than O_RDONLY mode.
4778 * systemd-rfkill has been reworked to become a singleton
4779 service that is activated through /dev/rfkill on each rfkill
4780 state change and saves the settings to disk. This way,
4781 systemd-rfkill is now compatible with devices that exist
4782 only intermittendly, and even restores state if the previous
4783 system shutdown was abrupt rather than clean.
4785 * The journal daemon gained support for vacuuming old journal
4786 files controlled by the number of files that shall remain,
4787 in addition to the already existing control by size and by
4788 date. This is useful as journal interleaving performance
4789 degrades with too many separate journal files, and allows
4790 putting an effective limit on them. The new setting defaults
4791 to 100, but this may be changed by setting SystemMaxFiles=
4792 and RuntimeMaxFiles= in journald.conf. Also, the
4793 "journalctl" tool gained the new --vacuum-files= switch to
4794 manually vacuum journal files to leave only the specified
4795 number of files in place.
4797 * udev will now create /dev/disk/by-path links for ATA devices
4798 on kernels where that is supported.
4800 * Galician, Serbian, Turkish and Korean translations were added.
4802 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
4803 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
4804 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
4805 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
4806 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
4807 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
4808 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
4809 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
4810 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
4811 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
4812 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
4813 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
4814 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
4815 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
4816 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
4817 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
4818 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
4819 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
4821 — Berlin, 2015-10-07
4825 * The DHCP implementation of systemd-networkd gained a set of
4828 - The DHCP server now supports emitting DNS and NTP
4829 information. It may be enabled and configured via
4830 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
4831 and NTP information is enabled, but no servers are
4832 configured, the corresponding uplink information (if there
4833 is any) is propagated.
4835 - Server and client now support transmission and reception
4836 of timezone information. It can be configured via the
4837 newly introduced network options UseTimezone=,
4838 EmitTimezone=, and Timezone=. Transmission of timezone
4839 information is enabled between host and containers by
4840 default now: the container will change its local timezone
4841 to what the host has set.
4843 - Lease timeouts can now be configured via
4844 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
4846 - The DHCP server improved on the stability of
4847 leases. Clients are more likely to get the same lease
4848 information back, even if the server loses state.
4850 - The DHCP server supports two new configuration options to
4851 control the lease address pool metrics, PoolOffset= and
4854 * The encapsulation limit of tunnels in systemd-networkd may
4855 now be configured via 'EncapsulationLimit='. It allows
4856 modifying the maximum additional levels of encapsulation
4857 that are permitted to be prepended to a packet.
4859 * systemd now supports the concept of user buses replacing
4860 session buses, if used with dbus-1.10 (and enabled via dbus
4861 --enable-user-session). It previously only supported this on
4862 kdbus-enabled systems, and this release expands this to
4863 'dbus-daemon' systems.
4865 * systemd-networkd now supports predictable interface names
4868 * systemd now optionally supports the new Linux kernel
4869 "unified" control group hierarchy. If enabled via the kernel
4870 command-line option 'systemd.unified_cgroup_hierarchy=1',
4871 systemd will try to mount the unified cgroup hierarchy
4872 directly on /sys/fs/cgroup. If not enabled, or not
4873 available, systemd will fall back to the legacy cgroup
4874 hierarchy setup, as before. Host system and containers can
4875 mix and match legacy and unified hierarchies as they
4876 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
4877 environment variable to individually select the hierarchy to
4878 use for executed containers. By default, nspawn will use the
4879 unified hierarchy for the containers if the host uses the
4880 unified hierarchy, and the legacy hierarchy otherwise.
4881 Please note that at this point the unified hierarchy is an
4882 experimental kernel feature and is likely to change in one
4883 of the next kernel releases. Therefore, it should not be
4884 enabled by default in downstream distributions yet. The
4885 minimum required kernel version for the unified hierarchy to
4886 work is 4.2. Note that when the unified hierarchy is used
4887 for the first time delegated access to controllers is
4888 safe. Because of this systemd-nspawn containers will get
4889 access to controllers now, as will systemd user
4890 sessions. This means containers and user sessions may now
4891 manage their own resources, partitioning up what the system
4894 * A new special scope unit "init.scope" has been introduced
4895 that encapsulates PID 1 of the system. It may be used to
4896 determine resource usage and enforce resource limits on PID
4897 1 itself. PID 1 hence moved out of the root of the control
4900 * The cgtop tool gained support for filtering out kernel
4901 threads when counting tasks in a control group. Also, the
4902 count of processes is now recursively summed up by
4903 default. Two options -k and --recursive= have been added to
4904 revert to old behaviour. The tool has also been updated to
4905 work correctly in containers now.
4907 * systemd-nspawn's --bind= and --bind-ro= options have been
4908 extended to allow creation of non-recursive bind mounts.
4910 * libsystemd gained two new calls sd_pid_get_cgroup() and
4911 sd_peer_get_cgroup() which return the control group path of
4912 a process or peer of a connected AF_UNIX socket. This
4913 function call is particularly useful when implementing
4914 delegated subtrees support in the control group hierarchy.
4916 * The "sd-event" event loop API of libsystemd now supports
4917 correct dequeuing of real-time signals, without losing
4920 * When systemd requests a polkit decision when managing units it
4921 will now add additional fields to the request, including unit
4922 name and desired operation. This enables more powerful polkit
4923 policies, that make decisions depending on these parameters.
4925 * nspawn learnt support for .nspawn settings files, that may
4926 accompany the image files or directories of containers, and
4927 may contain additional settings for the container. This is
4928 an alternative to configuring container parameters via the
4929 nspawn command line.
4931 Contributions from: Cristian Rodríguez, Daniel Mack, David
4932 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
4933 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
4934 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
4935 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
4936 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
4937 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
4938 Andersen, Tom Gundersen, Torstein Husebø
4940 — Berlin, 2015-09-08
4944 * machinectl gained a new verb 'shell' which opens a fresh
4945 shell on the target container or the host. It is similar to
4946 the existing 'login' command of machinectl, but spawns the
4947 shell directly without prompting for username or
4948 password. The pseudo machine '.host' now refers to the local
4949 host and is used by default. Hence, 'machinectl shell' can
4950 be used as replacement for 'su -' which spawns a session as
4951 a fresh systemd unit in a way that is fully isolated from
4952 the originating session.
4954 * systemd-networkd learned to cope with private-zone DHCP
4955 options and allows other programs to query the values.
4957 * SELinux access control when enabling/disabling units is no
4958 longer enforced with this release. The previous implementation
4959 was incorrect, and a new corrected implementation is not yet
4960 available. As unit file operations are still protected via
4961 polkit and D-Bus policy this is not a security problem. Yet,
4962 distributions which care about optimal SELinux support should
4963 probably not stabilize on this release.
4965 * sd-bus gained support for matches of type "arg0has=", that
4966 test for membership of strings in string arrays sent in bus
4969 * systemd-resolved now dumps the contents of its DNS and LLMNR
4970 caches to the logs on reception of the SIGUSR1 signal. This
4971 is useful to debug DNS behaviour.
4973 * The coredumpctl tool gained a new --directory= option to
4974 operate on journal files in a specific directory.
4976 * "systemctl reboot" and related commands gained a new
4977 "--message=" option which may be used to set a free-text
4978 wall message when shutting down or rebooting the
4979 system. This message is also logged, which is useful for
4980 figuring out the reason for a reboot or shutdown a
4983 * The "systemd-resolve-host" tool's -i switch now takes
4984 network interface numbers as alternative to interface names.
4986 * A new unit file setting for services has been introduced:
4987 UtmpMode= allows configuration of how precisely systemd
4988 handles utmp and wtmp entries for the service if this is
4989 enabled. This allows writing services that appear similar to
4990 user sessions in the output of the "w", "who", "last" and
4993 * systemd-resolved will now locally synthesize DNS resource
4994 records for the "localhost" and "gateway" domains as well as
4995 the local hostname. This should ensure that clients querying
4996 RRs via resolved will get similar results as those going via
4997 NSS, if nss-myhostname is enabled.
4999 Contributions from: Alastair Hughes, Alex Crawford, Daniel
5000 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
5001 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
5002 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
5003 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
5004 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
5005 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
5006 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
5007 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
5008 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
5009 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
5010 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
5012 — Berlin, 2015-08-27
5016 * The systemd-efi-boot-generator functionality was merged into
5017 systemd-gpt-auto-generator.
5019 * systemd-networkd now supports Group Policy for vxlan
5020 devices. It can be enabled via the new boolean configuration
5021 option called 'GroupPolicyExtension='.
5023 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
5024 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
5025 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
5027 — Berlin, 2015-07-31
5031 * The python-systemd code has been removed from the systemd repository.
5032 A new repository has been created which accommodates the code from
5033 now on, and we kindly ask distributions to create a separate package
5034 for this: https://github.com/systemd/python-systemd
5036 * The systemd daemon will now reload its main configuration
5037 (/etc/systemd/system.conf) on daemon-reload.
5039 * sd-dhcp now exposes vendor specific extensions via
5040 sd_dhcp_lease_get_vendor_specific().
5042 * systemd-networkd gained a number of new configuration options.
5044 - A new boolean configuration option for TAP devices called
5045 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
5046 device, thus allowing to send and receive GSO packets.
5048 - A new tunnel configuration option called 'CopyDSCP='.
5049 If enabled, the DSCP field of ip6 tunnels is copied into the
5050 decapsulated packet.
5052 - A set of boolean bridge configuration options were added.
5053 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
5054 and 'UnicastFlood=' are now parsed by networkd and applied to the
5055 respective bridge link device via the respective IFLA_BRPORT_*
5058 - A new string configuration option to override the hostname sent
5059 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
5060 is true, networkd will use the configured hostname instead of the
5061 system hostname when sending DHCP requests.
5063 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
5064 networkd will configure the IPv6 flow-label of the tunnel device
5065 according to RFC2460.
5067 - The 'macvtap' virtual network devices are now supported, similar to
5068 the already supported 'macvlan' devices.
5070 * systemd-resolved now implements RFC5452 to improve resilience against
5071 cache poisoning. Additionally, source port randomization is enabled
5072 by default to further protect against DNS spoofing attacks.
5074 * nss-mymachines now supports translating UIDs and GIDs of running
5075 containers with user-namespaces enabled. If a container 'foo'
5076 translates a host uid 'UID' to the container uid 'TUID', then
5077 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
5078 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
5079 mapped as 'vg-foo-TGID'.
5081 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
5082 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
5083 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
5084 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
5085 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
5086 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
5087 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
5088 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
5089 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
5090 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
5092 — Berlin, 2015-07-29
5096 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
5097 There are no known issues with current sysfs, and udev does not need
5098 or should be used to work around such bugs.
5100 * udev does no longer enable USB HID power management. Several reports
5101 indicate, that some devices cannot handle that setting.
5103 * The udev accelerometer helper was removed. The functionality
5104 is now fully included in iio-sensor-proxy. But this means,
5105 older iio-sensor-proxy versions will no longer provide
5106 accelerometer/orientation data with this systemd version.
5107 Please upgrade iio-sensor-proxy to version 1.0.
5109 * networkd gained a new configuration option IPv6PrivacyExtensions=
5110 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
5111 for Stateless Address") on selected networks.
5113 * For the sake of fewer build-time dependencies and less code in the
5114 main repository, the python bindings are about to be removed in the
5115 next release. A new repository has been created which accommodates
5116 the code from now on, and we kindly ask distributions to create a
5117 separate package for this. The removal will take place in v223.
5119 https://github.com/systemd/python-systemd
5121 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
5122 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
5123 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
5124 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
5125 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
5126 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
5127 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
5128 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
5129 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
5130 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
5132 — Berlin, 2015-07-07
5136 * The sd-bus.h and sd-event.h APIs have now been declared
5137 stable and have been added to the official interface of
5138 libsystemd.so. sd-bus implements an alternative D-Bus client
5139 library, that is relatively easy to use, very efficient and
5140 supports both classic D-Bus as well as kdbus as transport
5141 backend. sd-event is a generic event loop abstraction that
5142 is built around Linux epoll, but adds features such as event
5143 prioritization or efficient timer handling. Both APIs are good
5144 choices for C programs looking for a bus and/or event loop
5145 implementation that is minimal and does not have to be
5146 portable to other kernels.
5148 * kdbus support is no longer compile-time optional. It is now
5149 always built-in. However, it can still be disabled at
5150 runtime using the kdbus=0 kernel command line setting, and
5151 that setting may be changed to default to off, by specifying
5152 --disable-kdbus at build-time. Note though that the kernel
5153 command line setting has no effect if the kdbus.ko kernel
5154 module is not installed, in which case kdbus is (obviously)
5155 also disabled. We encourage all downstream distributions to
5156 begin testing kdbus by adding it to the kernel images in the
5157 development distributions, and leaving kdbus support in
5160 * The minimal required util-linux version has been bumped to
5163 * Support for chkconfig (--enable-chkconfig) was removed in
5164 favor of calling an abstraction tool
5165 /lib/systemd/systemd-sysv-install. This needs to be
5166 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
5167 in README for details.
5169 * If there's a systemd unit and a SysV init script for the
5170 same service name, and the user executes "systemctl enable"
5171 for it (or a related call), then this will now enable both
5172 (or execute the related operation on both), not just the
5175 * The libudev API documentation has been converted from gtkdoc
5178 * gudev has been removed from the systemd tree, it is now an
5181 * The systemd-cgtop tool learnt a new --raw switch to generate
5182 "raw" (machine parsable) output.
5184 * networkd's IPForwarding= .network file setting learnt the
5185 new setting "kernel", which ensures that networkd does not
5186 change the IP forwarding sysctl from the default kernel
5189 * The systemd-logind bus API now exposes a new boolean
5190 property "Docked" that reports whether logind considers the
5191 system "docked", i.e. connected to a docking station or not.
5193 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
5194 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
5195 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
5196 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
5197 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
5198 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
5199 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
5200 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
5201 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
5202 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
5203 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
5204 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
5205 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
5206 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
5207 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
5208 Fink, Zbigniew Jędrzejewski-Szmek
5210 — Berlin, 2015-06-19
5214 * The gudev library has been extracted into a separate repository
5215 available at: https://git.gnome.org/browse/libgudev/
5216 It is now managed as part of the Gnome project. Distributions
5217 are recommended to pass --disable-gudev to systemd and use
5218 gudev from the Gnome project instead. gudev is still included
5219 in systemd, for now. It will be removed soon, though. Please
5220 also see the announcement-thread on systemd-devel:
5221 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
5223 * systemd now exposes a CPUUsageNSec= property for each
5224 service unit on the bus, that contains the overall consumed
5225 CPU time of a service (the sum of what each process of the
5226 service consumed). This value is only available if
5227 CPUAccounting= is turned on for a service, and is then shown
5228 in the "systemctl status" output.
5230 * Support for configuring alternative mappings of the old SysV
5231 runlevels to systemd targets has been removed. They are now
5232 hardcoded in a way that runlevels 2, 3, 4 all map to
5233 multi-user.target and 5 to graphical.target (which
5234 previously was already the default behaviour).
5236 * The auto-mounter logic gained support for mount point
5237 expiry, using a new TimeoutIdleSec= setting in .automount
5238 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
5240 * The EFI System Partition (ESP) as mounted to /boot by
5241 systemd-efi-boot-generator will now be unmounted
5242 automatically after 2 minutes of not being used. This should
5243 minimize the risk of ESP corruptions.
5245 * New /etc/fstab options x-systemd.requires= and
5246 x-systemd.requires-mounts-for= are now supported to express
5247 additional dependencies for mounts. This is useful for
5248 journalling file systems that support external journal
5249 devices or overlay file systems that require underlying file
5250 systems to be mounted.
5252 * systemd does not support direct live-upgrades (via systemctl
5253 daemon-reexec) from versions older than v44 anymore. As no
5254 distribution we are aware of shipped such old versions in a
5255 stable release this should not be problematic.
5257 * When systemd forks off a new per-connection service instance
5258 it will now set the $REMOTE_ADDR environment variable to the
5259 remote IP address, and $REMOTE_PORT environment variable to
5260 the remote IP port. This behaviour is similar to the
5261 corresponding environment variables defined by CGI.
5263 * systemd-networkd gained support for uplink failure
5264 detection. The BindCarrier= option allows binding interface
5265 configuration dynamically to the link sense of other
5266 interfaces. This is useful to achieve behaviour like in
5269 * systemd-networkd gained support for configuring the DHCP
5270 client identifier to use when requesting leases.
5272 * systemd-networkd now has a per-network UseNTP= option to
5273 configure whether NTP server information acquired via DHCP
5274 is passed on to services like systemd-timesyncd.
5276 * systemd-networkd gained support for vti6 tunnels.
5278 * Note that systemd-networkd manages the sysctl variable
5279 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
5280 it is configured for since v219. The variable controls IP
5281 forwarding, and is a per-interface alternative to the global
5282 /proc/sys/net/ipv[46]/ip_forward. This setting is
5283 configurable in the IPForward= option, which defaults to
5284 "no". This means if networkd is used for an interface it is
5285 no longer sufficient to set the global sysctl option to turn
5286 on IP forwarding! Instead, the .network file option
5287 IPForward= needs to be turned on! Note that the
5288 implementation of this behaviour was broken in v219 and has
5291 * Many bonding and vxlan options are now configurable in
5294 * systemd-nspawn gained a new --property= setting to set unit
5295 properties for the container scope. This is useful for
5296 setting resource parameters (e.g. "CPUShares=500") on
5297 containers started from the command line.
5299 * systemd-nspawn gained a new --private-users= switch to make
5300 use of user namespacing available on recent Linux kernels.
5302 * systemd-nspawn may now be called as part of a shell pipeline
5303 in which case the pipes used for stdin and stdout are passed
5304 directly to the process invoked in the container, without
5305 indirection via a pseudo tty.
5307 * systemd-nspawn gained a new switch to control the UNIX
5308 signal to use when killing the init process of the container
5311 * systemd-nspawn gained a new --overlay= switch for mounting
5312 overlay file systems into the container using the new kernel
5315 * When a container image is imported via systemd-importd and
5316 the host file system is not btrfs, a loopback block device
5317 file is created in /var/lib/machines.raw with a btrfs file
5318 system inside. It is then mounted to /var/lib/machines to
5319 enable btrfs features for container management. The loopback
5320 file and btrfs file system is grown as needed when container
5321 images are imported via systemd-importd.
5323 * systemd-machined/systemd-importd gained support for btrfs
5324 quota, to enforce container disk space limits on disk. This
5325 is exposed in "machinectl set-limit".
5327 * systemd-importd now can import containers from local .tar,
5328 .raw and .qcow2 images, and export them to .tar and .raw. It
5329 can also import dkr v2 images now from the network (on top
5332 * systemd-importd gained support for verifying downloaded
5333 images with gpg2 (previously only gpg1 was supported).
5335 * systemd-machined, systemd-logind, systemd: most bus calls are
5336 now accessible to unprivileged processes via polkit. Also,
5337 systemd-logind will now allow users to kill their own sessions
5338 without further privileges or authorization.
5340 * systemd-shutdownd has been removed. This service was
5341 previously responsible for implementing scheduled shutdowns
5342 as exposed in /usr/bin/shutdown's time parameter. This
5343 functionality has now been moved into systemd-logind and is
5344 accessible via a bus interface.
5346 * "systemctl reboot" gained a new switch --firmware-setup that
5347 can be used to reboot into the EFI firmware setup, if that
5348 is available. systemd-logind now exposes an API on the bus
5349 to trigger such reboots, in case graphical desktop UIs want
5350 to cover this functionality.
5352 * "systemctl enable", "systemctl disable" and "systemctl mask"
5353 now support a new "--now" switch. If specified the units
5354 that are enabled will also be started, and the ones
5355 disabled/masked also stopped.
5357 * The Gummiboot EFI boot loader tool has been merged into
5358 systemd, and renamed to "systemd-boot". The bootctl tool has been
5359 updated to support systemd-boot.
5361 * An EFI kernel stub has been added that may be used to create
5362 kernel EFI binaries that contain not only the actual kernel,
5363 but also an initrd, boot splash, command line and OS release
5364 information. This combined binary can then be signed as a
5365 single image, so that the firmware can verify it all in one
5366 step. systemd-boot has special support for EFI binaries created
5367 like this and can extract OS release information from them
5368 and show them in the boot menu. This functionality is useful
5369 to implement cryptographically verified boot schemes.
5371 * Optional support has been added to systemd-fsck to pass
5372 fsck's progress report to an AF_UNIX socket in the file
5375 * udev will no longer create device symlinks for all block
5376 devices by default. A blacklist for excluding special block
5377 devices from this logic has been turned into a whitelist
5378 that requires picking block devices explicitly that require
5381 * A new (currently still internal) API sd-device.h has been
5382 added to libsystemd. This modernized API is supposed to
5383 replace libudev eventually. In fact, already much of libudev
5384 is now just a wrapper around sd-device.h.
5386 * A new hwdb database for storing metadata about pointing
5387 stick devices has been added.
5389 * systemd-tmpfiles gained support for setting file attributes
5390 similar to the "chattr" tool with new 'h' and 'H' lines.
5392 * systemd-journald will no longer unconditionally set the
5393 btrfs NOCOW flag on new journal files. This is instead done
5394 with tmpfiles snippet using the new 'h' line type. This
5395 allows easy disabling of this logic, by masking the
5396 journal-nocow.conf tmpfiles file.
5398 * systemd-journald will now translate audit message types to
5399 human readable identifiers when writing them to the
5400 journal. This should improve readability of audit messages.
5402 * The LUKS logic gained support for the offset= and skip=
5403 options in /etc/crypttab, as previously implemented by
5406 * /usr/lib/os-release gained a new optional field VARIANT= for
5407 distributions that support multiple variants (such as a
5408 desktop edition, a server edition, ...)
5410 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
5411 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
5412 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
5413 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
5414 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
5415 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
5416 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
5417 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
5418 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
5419 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
5420 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
5421 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
5422 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
5423 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
5424 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
5425 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
5426 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
5427 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
5428 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
5429 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
5430 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
5431 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
5432 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
5433 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
5434 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
5435 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
5436 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
5438 — Berlin, 2015-05-22
5442 * Introduce a new API "sd-hwdb.h" for querying the hardware
5443 metadata database. With this minimal interface one can query
5444 and enumerate the udev hwdb, decoupled from the old libudev
5445 library. libudev's interface for this is now only a wrapper
5446 around sd-hwdb. A new tool systemd-hwdb has been added to
5447 interface with and update the database.
5449 * When any of systemd's tools copies files (for example due to
5450 tmpfiles' C lines) a btrfs reflink will attempted first,
5451 before bytewise copying is done.
5453 * systemd-nspawn gained a new --ephemeral switch. When
5454 specified a btrfs snapshot is taken of the container's root
5455 directory, and immediately removed when the container
5456 terminates again. Thus, a container can be started whose
5457 changes never alter the container's root directory, and are
5458 lost on container termination. This switch can also be used
5459 for starting a container off the root file system of the
5460 host without affecting the host OS. This switch is only
5461 available on btrfs file systems.
5463 * systemd-nspawn gained a new --template= switch. It takes the
5464 path to a container tree to use as template for the tree
5465 specified via --directory=, should that directory be
5466 missing. This allows instantiating containers dynamically,
5467 on first run. This switch is only available on btrfs file
5470 * When a .mount unit refers to a mount point on which multiple
5471 mounts are stacked, and the .mount unit is stopped all of
5472 the stacked mount points will now be unmounted until no
5473 mount point remains.
5475 * systemd now has an explicit notion of supported and
5476 unsupported unit types. Jobs enqueued for unsupported unit
5477 types will now fail with an "unsupported" error code. More
5478 specifically .swap, .automount and .device units are not
5479 supported in containers, .busname units are not supported on
5480 non-kdbus systems. .swap and .automount are also not
5481 supported if their respective kernel compile time options
5484 * machinectl gained support for two new "copy-from" and
5485 "copy-to" commands for copying files from a running
5486 container to the host or vice versa.
5488 * machinectl gained support for a new "bind" command to bind
5489 mount host directories into local containers. This is
5490 currently only supported for nspawn containers.
5492 * networkd gained support for configuring bridge forwarding
5493 database entries (fdb) from .network files.
5495 * A new tiny daemon "systemd-importd" has been added that can
5496 download container images in tar, raw, qcow2 or dkr formats,
5497 and make them available locally in /var/lib/machines, so
5498 that they can run as nspawn containers. The daemon can GPG
5499 verify the downloads (not supported for dkr, since it has no
5500 provisions for verifying downloads). It will transparently
5501 decompress bz2, xz, gzip compressed downloads if necessary,
5502 and restore sparse files on disk. The daemon uses privilege
5503 separation to ensure the actual download logic runs with
5504 fewer privileges than the daemon itself. machinectl has
5505 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
5506 make the functionality of importd available to the
5507 user. With this in place the Fedora and Ubuntu "Cloud"
5508 images can be downloaded and booted as containers unmodified
5509 (the Fedora images lack the appropriate GPG signature files
5510 currently, so they cannot be verified, but this will change
5511 soon, hopefully). Note that downloading images is currently
5512 only fully supported on btrfs.
5514 * machinectl is now able to list container images found in
5515 /var/lib/machines, along with some metadata about sizes of
5516 disk and similar. If the directory is located on btrfs and
5517 quota is enabled, this includes quota display. A new command
5518 "image-status" has been added that shows additional
5519 information about images.
5521 * machinectl is now able to clone container images
5522 efficiently, if the underlying file system (btrfs) supports
5523 it, with the new "machinectl clone" command. It also
5524 gained commands for renaming and removing images, as well as
5525 marking them read-only or read-write (supported also on
5526 legacy file systems).
5528 * networkd gained support for collecting LLDP network
5529 announcements, from hardware that supports this. This is
5530 shown in networkctl output.
5532 * systemd-run gained support for a new -t (--pty) switch for
5533 invoking a binary on a pty whose input and output is
5534 connected to the invoking terminal. This allows executing
5535 processes as system services while interactively
5536 communicating with them via the terminal. Most interestingly
5537 this is supported across container boundaries. Invoking
5538 "systemd-run -t /bin/bash" is an alternative to running a
5539 full login session, the difference being that the former
5540 will not register a session, nor go through the PAM session
5543 * tmpfiles gained support for a new "v" line type for creating
5544 btrfs subvolumes. If the underlying file system is a legacy
5545 file system, this automatically degrades to creating a
5546 normal directory. Among others /var/lib/machines is now
5547 created like this at boot, should it be missing.
5549 * The directory /var/lib/containers/ has been deprecated and
5550 been replaced by /var/lib/machines. The term "machines" has
5551 been used in the systemd context as generic term for both
5552 VMs and containers, and hence appears more appropriate for
5553 this, as the directory can also contain raw images bootable
5556 * systemd-nspawn when invoked with -M but without --directory=
5557 or --image= is now capable of searching for the container
5558 root directory, subvolume or disk image automatically, in
5559 /var/lib/machines. systemd-nspawn@.service has been updated
5560 to make use of this, thus allowing it to be used for raw
5563 * A new machines.target unit has been introduced that is
5564 supposed to group all containers/VMs invoked as services on
5565 the system. systemd-nspawn@.service has been updated to
5566 integrate with that.
5568 * machinectl gained a new "start" command, for invoking a
5569 container as a service. "machinectl start foo" is mostly
5570 equivalent to "systemctl start systemd-nspawn@foo.service",
5571 but handles escaping in a nicer way.
5573 * systemd-nspawn will now mount most of the cgroupfs tree
5574 read-only into each container, with the exception of the
5575 container's own subtree in the name=systemd hierarchy.
5577 * journald now sets the special FS_NOCOW file flag for its
5578 journal files. This should improve performance on btrfs, by
5579 avoiding heavy fragmentation when journald's write-pattern
5580 is used on COW file systems. It degrades btrfs' data
5581 integrity guarantees for the files to the same levels as for
5582 ext3/ext4 however. This should be OK though as journald does
5583 its own data integrity checks and all its objects are
5584 checksummed on disk. Also, journald should handle btrfs disk
5585 full events a lot more gracefully now, by processing SIGBUS
5586 errors, and not relying on fallocate() anymore.
5588 * When journald detects that journal files it is writing to
5589 have been deleted it will immediately start new journal
5592 * systemd now provides a way to store file descriptors
5593 per-service in PID 1. This is useful for daemons to ensure
5594 that fds they require are not lost during a daemon
5595 restart. The fds are passed to the daemon on the next
5596 invocation in the same way socket activation fds are
5597 passed. This is now used by journald to ensure that the
5598 various sockets connected to all the system's stdout/stderr
5599 are not lost when journald is restarted. File descriptors
5600 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
5601 an extension to sd_notify(). Note that a limit is enforced
5602 on the number of fds a service can store in PID 1, and it
5603 defaults to 0, so that no fds may be stored, unless this is
5604 explicitly turned on.
5606 * The default TERM variable to use for units connected to a
5607 terminal, when no other value is explicitly is set is now
5608 vt220 rather than vt102. This should be fairly safe still,
5609 but allows PgUp/PgDn work.
5611 * The /etc/crypttab option header= as known from Debian is now
5614 * "loginctl user-status" and "loginctl session-status" will
5615 now show the last 10 lines of log messages of the
5616 user/session following the status output. Similar,
5617 "machinectl status" will show the last 10 log lines
5618 associated with a virtual machine or container
5619 service. (Note that this is usually not the log messages
5620 done in the VM/container itself, but simply what the
5621 container manager logs. For nspawn this includes all console
5624 * "loginctl session-status" without further argument will now
5625 show the status of the session of the caller. Similar,
5626 "lock-session", "unlock-session", "activate",
5627 "enable-linger", "disable-linger" may now be called without
5628 session/user parameter in which case they apply to the
5629 caller's session/user.
5631 * An X11 session scriptlet is now shipped that uploads
5632 $DISPLAY and $XAUTHORITY into the environment of the systemd
5633 --user daemon if a session begins. This should improve
5634 compatibility with X11 enabled applications run as systemd
5637 * Generators are now subject to masking via /etc and /run, the
5638 same way as unit files.
5640 * networkd .network files gained support for configuring
5641 per-link IPv4/IPv6 packet forwarding as well as IPv4
5642 masquerading. This is by default turned on for veth links to
5643 containers, as registered by systemd-nspawn. This means that
5644 nspawn containers run with --network-veth will now get
5645 automatic routed access to the host's networks without any
5646 further configuration or setup, as long as networkd runs on
5649 * systemd-nspawn gained the --port= (-p) switch to expose TCP
5650 or UDP posts of a container on the host. With this in place
5651 it is possible to run containers with private veth links
5652 (--network-veth), and have their functionality exposed on
5653 the host as if their services were running directly on the
5656 * systemd-nspawn's --network-veth switch now gained a short
5657 version "-n", since with the changes above it is now truly
5658 useful out-of-the-box. The systemd-nspawn@.service has been
5659 updated to make use of it too by default.
5661 * systemd-nspawn will now maintain a per-image R/W lock, to
5662 ensure that the same image is not started more than once
5663 writable. (It's OK to run an image multiple times
5664 simultaneously in read-only mode.)
5666 * systemd-nspawn's --image= option is now capable of
5667 dissecting and booting MBR and GPT disk images that contain
5668 only a single active Linux partition. Previously it
5669 supported only GPT disk images with proper GPT type
5670 IDs. This allows running cloud images from major
5671 distributions directly with systemd-nspawn, without
5674 * In addition to collecting mouse dpi data in the udev
5675 hardware database, there's now support for collecting angle
5676 information for mouse scroll wheels. The database is
5677 supposed to guarantee similar scrolling behavior on mice
5678 that it knows about. There's also support for collecting
5679 information about Touchpad types.
5681 * udev's input_id built-in will now also collect touch screen
5682 dimension data and attach it to probed devices.
5684 * /etc/os-release gained support for a Distribution Privacy
5687 * networkd gained support for creating "ipvlan", "gretap",
5688 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
5690 * systemd-tmpfiles gained support for "a" lines for setting
5693 * systemd-nspawn will now mount /tmp in the container to
5694 tmpfs, automatically.
5696 * systemd now exposes the memory.usage_in_bytes cgroup
5697 attribute and shows it for each service in the "systemctl
5698 status" output, if available.
5700 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
5701 immediate reboot is triggered. This useful if shutdown is
5702 hung and is unable to complete, to expedite the
5703 operation. Note that this kind of reboot will still unmount
5704 all file systems, and hence should not result in fsck being
5707 * A .device unit for an optical block device will now be
5708 considered active only when a medium is in the drive. Also,
5709 mount units are now bound to their backing devices thus
5710 triggering automatic unmounting when devices become
5711 unavailable. With this in place systemd will now
5712 automatically unmount left-over mounts when a CD-ROM is
5713 ejected or an USB stick is yanked from the system.
5715 * networkd-wait-online now has support for waiting for
5716 specific interfaces only (with globbing), and for giving up
5717 after a configurable timeout.
5719 * networkd now exits when idle. It will be automatically
5720 restarted as soon as interfaces show up, are removed or
5721 change state. networkd will stay around as long as there is
5722 at least one DHCP state machine or similar around, that keep
5725 * networkd may now configure IPv6 link-local addressing in
5726 addition to IPv4 link-local addressing.
5728 * The IPv6 "token" for use in SLAAC may now be configured for
5729 each .network interface in networkd.
5731 * Routes configured with networkd may now be assigned a scope
5734 * networkd's [Match] sections now support globbing and lists
5735 of multiple space-separated matches per item.
5737 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
5738 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
5739 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
5740 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
5741 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
5742 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
5743 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
5744 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
5745 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
5746 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
5747 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
5748 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
5749 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
5750 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
5751 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
5752 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
5753 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
5754 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
5755 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
5756 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
5757 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
5758 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
5759 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
5760 Hoffmann, Zbigniew Jędrzejewski-Szmek
5762 — Berlin, 2015-02-16
5766 * When querying unit file enablement status (for example via
5767 "systemctl is-enabled"), a new state "indirect" is now known
5768 which indicates that a unit might not be enabled itself, but
5769 another unit listed in its Also= setting might be.
5771 * Similar to the various existing ConditionXYZ= settings for
5772 units, there are now matching AssertXYZ= settings. While
5773 failing conditions cause a unit to be skipped, but its job
5774 to succeed, failing assertions declared like this will cause
5775 a unit start operation and its job to fail.
5777 * hostnamed now knows a new chassis type "embedded".
5779 * systemctl gained a new "edit" command. When used on a unit
5780 file, this allows extending unit files with .d/ drop-in
5781 configuration snippets or editing the full file (after
5782 copying it from /usr/lib to /etc). This will invoke the
5783 user's editor (as configured with $EDITOR), and reload the
5784 modified configuration after editing.
5786 * "systemctl status" now shows the suggested enablement state
5787 for a unit, as declared in the (usually vendor-supplied)
5788 system preset files.
5790 * nss-myhostname will now resolve the single-label host name
5791 "gateway" to the locally configured default IP routing
5792 gateways, ordered by their metrics. This assigns a stable
5793 name to the used gateways, regardless which ones are
5794 currently configured. Note that the name will only be
5795 resolved after all other name sources (if nss-myhostname is
5796 configured properly) and should hence not negatively impact
5797 systems that use the single-label host name "gateway" in
5800 * systemd-inhibit now allows filtering by mode when listing
5803 * Scope and service units gained a new "Delegate" boolean
5804 property, which, when set, allows processes running inside the
5805 unit to further partition resources. This is primarily
5806 useful for systemd user instances as well as container
5809 * journald will now pick up audit messages directly from
5810 the kernel, and log them like any other log message. The
5811 audit fields are split up and fully indexed. This means that
5812 journalctl in many ways is now a (nicer!) alternative to
5813 ausearch, the traditional audit client. Note that this
5814 implements only a minimal audit client. If you want the
5815 special audit modes like reboot-on-log-overflow, please use
5816 the traditional auditd instead, which can be used in
5817 parallel to journald.
5819 * The ConditionSecurity= unit file option now understands the
5820 special string "audit" to check whether auditing is
5823 * journalctl gained two new commands --vacuum-size= and
5824 --vacuum-time= to delete old journal files until the
5825 remaining ones take up no more than the specified size on disk,
5826 or are not older than the specified time.
5828 * A new, native PPPoE library has been added to sd-network,
5829 systemd's library of light-weight networking protocols. This
5830 library will be used in a future version of networkd to
5831 enable PPPoE communication without an external pppd daemon.
5833 * The busctl tool now understands a new "capture" verb that
5834 works similar to "monitor", but writes a packet capture
5835 trace to STDOUT that can be redirected to a file which is
5836 compatible with libcap's capture file format. This can then
5837 be loaded in Wireshark and similar tools to inspect bus
5840 * The busctl tool now understands a new "tree" verb that shows
5841 the object trees of a specific service on the bus, or of all
5844 * The busctl tool now understands a new "introspect" verb that
5845 shows all interfaces and members of objects on the bus,
5846 including their signature and values. This is particularly
5847 useful to get more information about bus objects shown by
5848 the new "busctl tree" command.
5850 * The busctl tool now understands new verbs "call",
5851 "set-property" and "get-property" for invoking bus method
5852 calls, setting and getting bus object properties in a
5855 * busctl gained a new --augment-creds= argument that controls
5856 whether the tool shall augment credential information it
5857 gets from the bus with data from /proc, in a possibly
5860 * nspawn's --link-journal= switch gained two new values
5861 "try-guest" and "try-host" that work like "guest" and
5862 "host", but do not fail if the host has no persistent
5863 journalling enabled. -j is now equivalent to
5864 --link-journal=try-guest.
5866 * macvlan network devices created by nspawn will now have
5867 stable MAC addresses.
5869 * A new SmackProcessLabel= unit setting has been added, which
5870 controls the SMACK security label processes forked off by
5871 the respective unit shall use.
5873 * If compiled with --enable-xkbcommon, systemd-localed will
5874 verify x11 keymap settings by compiling the given keymap. It
5875 will spew out warnings if the compilation fails. This
5876 requires libxkbcommon to be installed.
5878 * When a coredump is collected, a larger number of metadata
5879 fields is now collected and included in the journal records
5880 created for it. More specifically, control group membership,
5881 environment variables, memory maps, working directory,
5882 chroot directory, /proc/$PID/status, and a list of open file
5883 descriptors is now stored in the log entry.
5885 * The udev hwdb now contains DPI information for mice. For
5888 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
5890 * All systemd programs that read standalone configuration
5891 files in /etc now also support a corresponding series of
5892 .conf.d configuration directories in /etc/, /run/,
5893 /usr/local/lib/, /usr/lib/, and (if configured with
5894 --enable-split-usr) /lib/. In particular, the following
5895 configuration files now have corresponding configuration
5896 directories: system.conf user.conf, logind.conf,
5897 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
5898 resolved.conf, timesyncd.conf, journal-remote.conf, and
5899 journal-upload.conf. Note that distributions should use the
5900 configuration directories in /usr/lib/; the directories in
5901 /etc/ are reserved for the system administrator.
5903 * systemd-rfkill will no longer take the rfkill device name
5904 into account when storing rfkill state on disk, as the name
5905 might be dynamically assigned and not stable. Instead, the
5906 ID_PATH udev variable combined with the rfkill type (wlan,
5907 bluetooth, ...) is used.
5909 * A new service systemd-machine-id-commit.service has been
5910 added. When used on systems where /etc is read-only during
5911 boot, and /etc/machine-id is not initialized (but an empty
5912 file), this service will copy the temporary machine ID
5913 created as replacement into /etc after the system is fully
5914 booted up. This is useful for systems that are freshly
5915 installed with a non-initialized machine ID, but should get
5916 a fixed machine ID for subsequent boots.
5918 * networkd's .netdev files now provide a large set of
5919 configuration parameters for VXLAN devices. Similarly, the
5920 bridge port cost parameter is now configurable in .network
5921 files. There's also new support for configuring IP source
5922 routing. networkd .link files gained support for a new
5923 OriginalName= match that is useful to match against the
5924 original interface name the kernel assigned. .network files
5925 may include MTU= and MACAddress= fields for altering the MTU
5926 and MAC address while being connected to a specific network
5929 * The LUKS logic gained supported for configuring
5930 UUID-specific key files. There's also new support for naming
5931 LUKS device from the kernel command line, using the new
5932 luks.name= argument.
5934 * Timer units may now be transiently created via the bus API
5935 (this was previously already available for scope and service
5936 units). In addition it is now possible to create multiple
5937 transient units at the same time with a single bus call. The
5938 "systemd-run" tool has been updated to make use of this for
5939 running commands on a specified time, in at(1)-style.
5941 * tmpfiles gained support for "t" lines, for assigning
5942 extended attributes to files. Among other uses this may be
5943 used to assign SMACK labels to files.
5945 Contributions from: Alin Rauta, Alison Chaiken, Andrej
5946 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
5947 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
5948 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
5949 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
5950 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
5951 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
5952 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
5953 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
5954 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
5955 Michael Biebl, Michael Chapman, Michael Marineau, Michal
5956 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
5957 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
5958 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
5959 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
5960 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5961 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
5962 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
5964 — Berlin, 2014-12-10
5968 * journalctl gained the new options -t/--identifier= to match
5969 on the syslog identifier (aka "tag"), as well as --utc to
5970 show log timestamps in the UTC timezone. journalctl now also
5971 accepts -n/--lines=all to disable line capping in a pager.
5973 * journalctl gained a new switch, --flush, that synchronously
5974 flushes logs from /run/log/journal to /var/log/journal if
5975 persistent storage is enabled. systemd-journal-flush.service
5976 now waits until the operation is complete.
5978 * Services can notify the manager before they start a reload
5979 (by sending RELOADING=1) or shutdown (by sending
5980 STOPPING=1). This allows the manager to track and show the
5981 internal state of daemons and closes a race condition when
5982 the process is still running but has closed its D-Bus
5985 * Services with Type=oneshot do not have to have any ExecStart
5988 * User units are now loaded also from
5989 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
5990 /run/systemd/user directory that was already previously
5991 supported, but is under the control of the user.
5993 * Job timeouts (i.e. timeouts on the time a job that is
5994 queued stays in the run queue) can now optionally result in
5995 immediate reboot or power-off actions (JobTimeoutAction= and
5996 JobTimeoutRebootArgument=). This is useful on ".target"
5997 units, to limit the maximum time a target remains
5998 undispatched in the run queue, and to trigger an emergency
5999 operation in such a case. This is now used by default to
6000 turn off the system if boot-up (as defined by everything in
6001 basic.target) hangs and does not complete for at least
6002 15min. Also, if power-off or reboot hang for at least 30min
6003 an immediate power-off/reboot operation is triggered. This
6004 functionality is particularly useful to increase reliability
6005 on embedded devices, but also on laptops which might
6006 accidentally get powered on when carried in a backpack and
6007 whose boot stays stuck in a hard disk encryption passphrase
6010 * systemd-logind can be configured to also handle lid switch
6011 events even when the machine is docked or multiple displays
6012 are attached (HandleLidSwitchDocked= option).
6014 * A helper binary and a service have been added which can be
6015 used to resume from hibernation in the initramfs. A
6016 generator will parse the resume= option on the kernel
6017 command line to trigger resume.
6019 * A user console daemon systemd-consoled has been
6020 added. Currently, it is a preview, and will so far open a
6021 single terminal on each session of the user marked as
6022 Desktop=systemd-console.
6024 * Route metrics can be specified for DHCP routes added by
6027 * The SELinux context of socket-activated services can be set
6028 from the information provided by the networking stack
6029 (SELinuxContextFromNet= option).
6031 * Userspace firmware loading support has been removed and
6032 the minimum supported kernel version is thus bumped to 3.7.
6034 * Timeout for udev workers has been increased from 1 to 3
6035 minutes, but a warning will be printed after 1 minute to
6036 help diagnose kernel modules that take a long time to load.
6038 * Udev rules can now remove tags on devices with TAG-="foobar".
6040 * systemd's readahead implementation has been removed. In many
6041 circumstances it didn't give expected benefits even for
6042 rotational disk drives and was becoming less relevant in the
6043 age of SSDs. As none of the developers has been using
6044 rotating media anymore, and nobody stepped up to actively
6045 maintain this component of systemd it has now been removed.
6047 * Swap units can use Options= to specify discard options.
6048 Discard options specified for swaps in /etc/fstab are now
6051 * Docker containers are now detected as a separate type of
6054 * The Password Agent protocol gained support for queries where
6055 the user input is shown, useful e.g. for user names.
6056 systemd-ask-password gained a new --echo option to turn that
6059 * The default sysctl.d/ snippets will now set:
6061 net.core.default_qdisc = fq_codel
6063 This selects Fair Queuing Controlled Delay as the default
6064 queuing discipline for network interfaces. fq_codel helps
6065 fight the network bufferbloat problem. It is believed to be
6066 a good default with no tuning required for most workloads.
6067 Downstream distributions may override this choice. On 10Gbit
6068 servers that do not do forwarding, "fq" may perform better.
6069 Systems without a good clocksource should use "pfifo_fast".
6071 * If kdbus is enabled during build a new option BusPolicy= is
6072 available for service units, that allows locking all service
6073 processes into a stricter bus policy, in order to limit
6074 access to various bus services, or even hide most of them
6075 from the service's view entirely.
6077 * networkctl will now show the .network and .link file
6078 networkd has applied to a specific interface.
6080 * sd-login gained a new API call sd_session_get_desktop() to
6081 query which desktop environment has been selected for a
6084 * UNIX utmp support is now compile-time optional to support
6085 legacy-free systems.
6087 * systemctl gained two new commands "add-wants" and
6088 "add-requires" for pulling in units from specific targets
6091 * If the word "rescue" is specified on the kernel command line
6092 the system will now boot into rescue mode (aka
6093 rescue.target), which was previously available only by
6094 specifying "1" or "systemd.unit=rescue.target" on the kernel
6095 command line. This new kernel command line option nicely
6096 mirrors the already existing "emergency" kernel command line
6099 * New kernel command line options mount.usr=, mount.usrflags=,
6100 mount.usrfstype= have been added that match root=, rootflags=,
6101 rootfstype= but allow mounting a specific file system to
6104 * The $NOTIFY_SOCKET is now also passed to control processes of
6105 services, not only the main process.
6107 * This version reenables support for fsck's -l switch. This
6108 means at least version v2.25 of util-linux is required for
6109 operation, otherwise dead-locks on device nodes may
6110 occur. Again: you need to update util-linux to at least
6111 v2.25 when updating systemd to v217.
6113 * The "multi-seat-x" tool has been removed from systemd, as
6114 its functionality has been integrated into X servers 1.16,
6115 and the tool is hence redundant. It is recommended to update
6116 display managers invoking this tool to simply invoke X
6117 directly from now on, again.
6119 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
6120 message flag has been added for all of systemd's polkit
6121 authenticated method calls has been added. In particular this
6122 now allows optional interactive authorization via polkit for
6123 many of PID1's privileged operations such as unit file
6124 enabling and disabling.
6126 * "udevadm hwdb --update" learnt a new switch "--usr" for
6127 placing the rebuilt hardware database in /usr instead of
6128 /etc. When used only hardware database entries stored in
6129 /usr will be used, and any user database entries in /etc are
6130 ignored. This functionality is useful for vendors to ship a
6131 pre-built database on systems where local configuration is
6132 unnecessary or unlikely.
6134 * Calendar time specifications in .timer units now also
6135 understand the strings "semi-annually", "quarterly" and
6136 "minutely" as shortcuts (in addition to the preexisting
6137 "anually", "hourly", ...).
6139 * systemd-tmpfiles will now correctly create files in /dev
6140 at boot which are marked for creation only at boot. It is
6141 recommended to always create static device nodes with 'c!'
6142 and 'b!', so that they are created only at boot and not
6143 overwritten at runtime.
6145 * When the watchdog logic is used for a service (WatchdogSec=)
6146 and the watchdog timeout is hit the service will now be
6147 terminated with SIGABRT (instead of just SIGTERM), in order
6148 to make sure a proper coredump and backtrace is
6149 generated. This ensures that hanging services will result in
6150 similar coredump/backtrace behaviour as services that hit a
6153 Contributions from: Andreas Henriksson, Andrei Borzenkov,
6154 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
6155 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
6156 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
6157 Herrmann, David Sommerseth, David Strauss, Emil Renner
6158 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
6159 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
6160 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
6161 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
6162 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
6163 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
6164 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
6165 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
6166 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
6167 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
6168 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
6169 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
6170 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
6171 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
6172 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
6173 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
6176 — Berlin, 2014-10-28
6180 * timedated no longer reads NTP implementation unit names from
6181 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
6182 implementations should add a
6184 Conflicts=systemd-timesyncd.service
6186 to their unit files to take over and replace systemd's NTP
6187 default functionality.
6189 * systemd-sysusers gained a new line type "r" for configuring
6190 which UID/GID ranges to allocate system users/groups
6191 from. Lines of type "u" may now add an additional column
6192 that specifies the home directory for the system user to be
6193 created. Also, systemd-sysusers may now optionally read user
6194 information from STDIN instead of a file. This is useful for
6195 invoking it from RPM preinst scriptlets that need to create
6196 users before the first RPM file is installed since these
6197 files might need to be owned by them. A new
6198 %sysusers_create_inline RPM macro has been introduced to do
6199 just that. systemd-sysusers now updates the shadow files as
6200 well as the user/group databases, which should enhance
6201 compatibility with certain tools like grpck.
6203 * A number of bus APIs of PID 1 now optionally consult polkit to
6204 permit access for otherwise unprivileged clients under certain
6205 conditions. Note that this currently doesn't support
6206 interactive authentication yet, but this is expected to be
6207 added eventually, too.
6209 * /etc/machine-info now has new fields for configuring the
6210 deployment environment of the machine, as well as the
6211 location of the machine. hostnamectl has been updated with
6212 new command to update these fields.
6214 * systemd-timesyncd has been updated to automatically acquire
6215 NTP server information from systemd-networkd, which might
6216 have been discovered via DHCP.
6218 * systemd-resolved now includes a caching DNS stub resolver
6219 and a complete LLMNR name resolution implementation. A new
6220 NSS module "nss-resolve" has been added which can be used
6221 instead of glibc's own "nss-dns" to resolve hostnames via
6222 systemd-resolved. Hostnames, addresses and arbitrary RRs may
6223 be resolved via systemd-resolved D-Bus APIs. In contrast to
6224 the glibc internal resolver systemd-resolved is aware of
6225 multi-homed system, and keeps DNS server and caches separate
6226 and per-interface. Queries are sent simultaneously on all
6227 interfaces that have DNS servers configured, in order to
6228 properly handle VPNs and local LANs which might resolve
6229 separate sets of domain names. systemd-resolved may acquire
6230 DNS server information from systemd-networkd automatically,
6231 which in turn might have discovered them via DHCP. A tool
6232 "systemd-resolve-host" has been added that may be used to
6233 query the DNS logic in resolved. systemd-resolved implements
6234 IDNA and automatically uses IDNA or UTF-8 encoding depending
6235 on whether classic DNS or LLMNR is used as transport. In the
6236 next releases we intend to add a DNSSEC and mDNS/DNS-SD
6237 implementation to systemd-resolved.
6239 * A new NSS module nss-mymachines has been added, that
6240 automatically resolves the names of all local registered
6241 containers to their respective IP addresses.
6243 * A new client tool "networkctl" for systemd-networkd has been
6244 added. It currently is entirely passive and will query
6245 networking configuration from udev, rtnetlink and networkd,
6246 and present it to the user in a very friendly
6247 way. Eventually, we hope to extend it to become a full
6248 control utility for networkd.
6250 * .socket units gained a new DeferAcceptSec= setting that
6251 controls the kernels' TCP_DEFER_ACCEPT sockopt for
6252 TCP. Similarly, support for controlling TCP keep-alive
6253 settings has been added (KeepAliveTimeSec=,
6254 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
6255 turning off Nagle's algorithm on TCP has been added
6258 * logind learned a new session type "web", for use in projects
6259 like Cockpit which register web clients as PAM sessions.
6261 * timer units with at least one OnCalendar= setting will now
6262 be started only after time-sync.target has been
6263 reached. This way they will not elapse before the system
6264 clock has been corrected by a local NTP client or
6265 similar. This is particular useful on RTC-less embedded
6266 machines, that come up with an invalid system clock.
6268 * systemd-nspawn's --network-veth= switch should now result in
6269 stable MAC addresses for both the outer and the inner side
6272 * systemd-nspawn gained a new --volatile= switch for running
6273 container instances with /etc or /var unpopulated.
6275 * The kdbus client code has been updated to use the new Linux
6276 3.17 memfd subsystem instead of the old kdbus-specific one.
6278 * systemd-networkd's DHCP client and server now support
6279 FORCERENEW. There are also new configuration options to
6280 configure the vendor client identifier and broadcast mode
6283 * systemd will no longer inform the kernel about the current
6284 timezone, as this is necessarily incorrect and racy as the
6285 kernel has no understanding of DST and similar
6286 concepts. This hence means FAT timestamps will be always
6287 considered UTC, similar to what Android is already
6288 doing. Also, when the RTC is configured to the local time
6289 (rather than UTC) systemd will never synchronize back to it,
6290 as this might confuse Windows at a later boot.
6292 * systemd-analyze gained a new command "verify" for offline
6293 validation of unit files.
6295 * systemd-networkd gained support for a couple of additional
6296 settings for bonding networking setups. Also, the metric for
6297 statically configured routes may now be configured. For
6298 network interfaces where this is appropriate the peer IP
6299 address may now be configured.
6301 * systemd-networkd's DHCP client will no longer request
6302 broadcasting by default, as this tripped up some networks.
6303 For hardware where broadcast is required the feature should
6304 be switched back on using RequestBroadcast=yes.
6306 * systemd-networkd will now set up IPv4LL addresses (when
6307 enabled) even if DHCP is configured successfully.
6309 * udev will now default to respect network device names given
6310 by the kernel when the kernel indicates that these are
6311 predictable. This behavior can be tweaked by changing
6312 NamePolicy= in the relevant .link file.
6314 * A new library systemd-terminal has been added that
6315 implements full TTY stream parsing and rendering. This
6316 library is supposed to be used later on for implementing a
6317 full userspace VT subsystem, replacing the current kernel
6320 * A new tool systemd-journal-upload has been added to push
6321 journal data to a remote system running
6322 systemd-journal-remote.
6324 * journald will no longer forward all local data to another
6325 running syslog daemon. This change has been made because
6326 rsyslog (which appears to be the most commonly used syslog
6327 implementation these days) no longer makes use of this, and
6328 instead pulls the data out of the journal on its own. Since
6329 forwarding the messages to a non-existent syslog server is
6330 more expensive than we assumed we have now turned this
6331 off. If you run a syslog server that is not a recent rsyslog
6332 version, you have to turn this option on again
6333 (ForwardToSyslog= in journald.conf).
6335 * journald now optionally supports the LZ4 compressor for
6336 larger journal fields. This compressor should perform much
6337 better than XZ which was the previous default.
6339 * machinectl now shows the IP addresses of local containers,
6340 if it knows them, plus the interface name of the container.
6342 * A new tool "systemd-escape" has been added that makes it
6343 easy to escape strings to build unit names and similar.
6345 * sd_notify() messages may now include a new ERRNO= field
6346 which is parsed and collected by systemd and shown among the
6347 "systemctl status" output for a service.
6349 * A new component "systemd-firstboot" has been added that
6350 queries the most basic systemd information (timezone,
6351 hostname, root password) interactively on first
6352 boot. Alternatively it may also be used to provision these
6353 things offline on OS images installed into directories.
6355 * The default sysctl.d/ snippets will now set
6357 net.ipv4.conf.default.promote_secondaries=1
6359 This has the benefit of no flushing secondary IP addresses
6360 when primary addresses are removed.
6362 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
6363 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
6364 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
6365 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
6366 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
6367 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
6368 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
6369 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
6370 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
6371 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
6372 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
6373 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
6374 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
6375 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
6376 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
6378 — Berlin, 2014-08-19
6382 * A new tool systemd-sysusers has been added. This tool
6383 creates system users and groups in /etc/passwd and
6384 /etc/group, based on static declarative system user/group
6385 definitions in /usr/lib/sysusers.d/. This is useful to
6386 enable factory resets and volatile systems that boot up with
6387 an empty /etc directory, and thus need system users and
6388 groups created during early boot. systemd now also ships
6389 with two default sysusers.d/ files for the most basic
6390 users and groups systemd and the core operating system
6393 * A new tmpfiles snippet has been added that rebuilds the
6394 essential files in /etc on boot, should they be missing.
6396 * A directive for ensuring automatic clean-up of
6397 /var/cache/man/ has been removed from the default
6398 configuration. This line should now be shipped by the man
6399 implementation. The necessary change has been made to the
6400 man-db implementation. Note that you need to update your man
6401 implementation to one that ships this line, otherwise no
6402 automatic clean-up of /var/cache/man will take place.
6404 * A new condition ConditionNeedsUpdate= has been added that
6405 may conditionalize services to only run when /etc or /var
6406 are "older" than the vendor operating system resources in
6407 /usr. This is useful for reconstructing or updating /etc
6408 after an offline update of /usr or a factory reset, on the
6409 next reboot. Services that want to run once after such an
6410 update or reset should use this condition and order
6411 themselves before the new systemd-update-done.service, which
6412 will mark the two directories as fully updated. A number of
6413 service files have been added making use of this, to rebuild
6414 the udev hardware database, the journald message catalog and
6415 dynamic loader cache (ldconfig). The systemd-sysusers tool
6416 described above also makes use of this now. With this in
6417 place it is now possible to start up a minimal operating
6418 system with /etc empty cleanly. For more information on the
6419 concepts involved see this recent blog story:
6421 http://0pointer.de/blog/projects/stateless.html
6423 * A new system group "input" has been introduced, and all
6424 input device nodes get this group assigned. This is useful
6425 for system-level software to get access to input devices. It
6426 complements what is already done for "audio" and "video".
6428 * systemd-networkd learnt minimal DHCPv4 server support in
6429 addition to the existing DHCPv4 client support. It also
6430 learnt DHCPv6 client and IPv6 Router Solicitation client
6431 support. The DHCPv4 client gained support for static routes
6432 passed in from the server. Note that the [DHCPv4] section
6433 known in older systemd-networkd versions has been renamed to
6434 [DHCP] and is now also used by the DHCPv6 client. Existing
6435 .network files using settings of this section should be
6436 updated, though compatibility is maintained. Optionally, the
6437 client hostname may now be sent to the DHCP server.
6439 * networkd gained support for vxlan virtual networks as well
6440 as tun/tap and dummy devices.
6442 * networkd gained support for automatic allocation of address
6443 ranges for interfaces from a system-wide pool of
6444 addresses. This is useful for dynamically managing a large
6445 number of interfaces with a single network configuration
6446 file. In particular this is useful to easily assign
6447 appropriate IP addresses to the veth links of a large number
6448 of nspawn instances.
6450 * RPM macros for processing sysusers, sysctl and binfmt
6451 drop-in snippets at package installation time have been
6454 * The /etc/os-release file should now be placed in
6455 /usr/lib/os-release. The old location is automatically
6456 created as symlink. /usr/lib is the more appropriate
6457 location of this file, since it shall actually describe the
6458 vendor operating system shipped in /usr, and not the
6459 configuration stored in /etc.
6461 * .mount units gained a new boolean SloppyOptions= setting
6462 that maps to mount(8)'s -s option which enables permissive
6463 parsing of unknown mount options.
6465 * tmpfiles learnt a new "L+" directive which creates a symlink
6466 but (unlike "L") deletes a pre-existing file first, should
6467 it already exist and not already be the correct
6468 symlink. Similarly, "b+", "c+" and "p+" directives have been
6469 added as well, which create block and character devices, as
6470 well as fifos in the filesystem, possibly removing any
6471 pre-existing files of different types.
6473 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
6474 'argument' field (which so far specified the source to
6475 symlink/copy the files from) is now optional. If omitted the
6476 same file os copied from /usr/share/factory/ suffixed by the
6477 full destination path. This is useful for populating /etc
6478 with essential files, by copying them from vendor defaults
6479 shipped in /usr/share/factory/etc.
6481 * A new command "systemctl preset-all" has been added that
6482 applies the service preset settings to all installed unit
6483 files. A new switch --preset-mode= has been added that
6484 controls whether only enable or only disable operations
6487 * A new command "systemctl is-system-running" has been added
6488 that allows checking the overall state of the system, for
6489 example whether it is fully up and running.
6491 * When the system boots up with an empty /etc, the equivalent
6492 to "systemctl preset-all" is executed during early boot, to
6493 make sure all default services are enabled after a factory
6496 * systemd now contains a minimal preset file that enables the
6497 most basic services systemd ships by default.
6499 * Unit files' [Install] section gained a new DefaultInstance=
6500 field for defining the default instance to create if a
6501 template unit is enabled with no instance specified.
6503 * A new passive target cryptsetup-pre.target has been added
6504 that may be used by services that need to make they run and
6505 finish before the first LUKS cryptographic device is set up.
6507 * The /dev/loop-control and /dev/btrfs-control device nodes
6508 are now owned by the "disk" group by default, opening up
6509 access to this group.
6511 * systemd-coredump will now automatically generate a
6512 stack trace of all core dumps taking place on the system,
6513 based on elfutils' libdw library. This stack trace is logged
6516 * systemd-coredump may now optionally store coredumps directly
6517 on disk (in /var/lib/systemd/coredump, possibly compressed),
6518 instead of storing them unconditionally in the journal. This
6519 mode is the new default. A new configuration file
6520 /etc/systemd/coredump.conf has been added to configure this
6521 and other parameters of systemd-coredump.
6523 * coredumpctl gained a new "info" verb to show details about a
6524 specific coredump. A new switch "-1" has also been added
6525 that makes sure to only show information about the most
6526 recent entry instead of all entries. Also, as the tool is
6527 generally useful now the "systemd-" prefix of the binary
6528 name has been removed. Distributions that want to maintain
6529 compatibility with the old name should add a symlink from
6530 the old name to the new name.
6532 * journald's SplitMode= now defaults to "uid". This makes sure
6533 that unprivileged users can access their own coredumps with
6534 coredumpctl without restrictions.
6536 * New kernel command line options "systemd.wants=" (for
6537 pulling an additional unit during boot), "systemd.mask="
6538 (for masking a specific unit for the boot), and
6539 "systemd.debug-shell" (for enabling the debug shell on tty9)
6540 have been added. This is implemented in the new generator
6541 "systemd-debug-generator".
6543 * systemd-nspawn will now by default filter a couple of
6544 syscalls for containers, among them those required for
6545 kernel module loading, direct x86 IO port access, swap
6546 management, and kexec. Most importantly though
6547 open_by_handle_at() is now prohibited for containers,
6548 closing a hole similar to a recently discussed vulnerability
6549 in docker regarding access to files on file hierarchies the
6550 container should normally not have access to. Note that, for
6551 nspawn, we generally make no security claims anyway (and
6552 this is explicitly documented in the man page), so this is
6553 just a fix for one of the most obvious problems.
6555 * A new man page file-hierarchy(7) has been added that
6556 contains a minimized, modernized version of the file system
6557 layout systemd expects, similar in style to the FHS
6558 specification or hier(5). A new tool systemd-path(1) has
6559 been added to query many of these paths for the local
6562 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
6563 longer done. Since the directory now has a per-user size
6564 limit, and is cleaned on logout this appears unnecessary,
6565 in particular since this now brings the lifecycle of this
6566 directory closer in line with how IPC objects are handled.
6568 * systemd.pc now exports a number of additional directories,
6569 including $libdir (which is useful to identify the library
6570 path for the primary architecture of the system), and a
6571 couple of drop-in directories.
6573 * udev's predictable network interface names now use the dev_port
6574 sysfs attribute, introduced in linux 3.15 instead of dev_id to
6575 distinguish between ports of the same PCI function. dev_id should
6576 only be used for ports using the same HW address, hence the need
6579 * machined has been updated to export the OS version of a
6580 container (read from /etc/os-release and
6581 /usr/lib/os-release) on the bus. This is now shown in
6582 "machinectl status" for a machine.
6584 * A new service setting RestartForceExitStatus= has been
6585 added. If configured to a set of exit signals or process
6586 return values, the service will be restarted when the main
6587 daemon process exits with any of them, regardless of the
6590 * systemctl's -H switch for connecting to remote systemd
6591 machines has been extended so that it may be used to
6592 directly connect to a specific container on the
6593 host. "systemctl -H root@foobar:waldi" will now connect as
6594 user "root" to host "foobar", and then proceed directly to
6595 the container named "waldi". Note that currently you have to
6596 authenticate as user "root" for this to work, as entering
6597 containers is a privileged operation.
6599 Contributions from: Andreas Henriksson, Benjamin Steinwender,
6600 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
6601 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
6602 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
6603 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
6604 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
6605 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
6606 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
6607 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
6608 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
6609 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
6610 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
6612 — Berlin, 2014-07-03
6616 * As an experimental feature, udev now tries to lock the
6617 disk device node (flock(LOCK_SH|LOCK_NB)) while it
6618 executes events for the disk or any of its partitions.
6619 Applications like partitioning programs can lock the
6620 disk device node (flock(LOCK_EX)) and claim temporary
6621 device ownership that way; udev will entirely skip all event
6622 handling for this disk and its partitions. If the disk
6623 was opened for writing, the close will trigger a partition
6624 table rescan in udev's "watch" facility, and if needed
6625 synthesize "change" events for the disk and all its partitions.
6626 This is now unconditionally enabled, and if it turns out to
6627 cause major problems, we might turn it on only for specific
6628 devices, or might need to disable it entirely. Device Mapper
6629 devices are excluded from this logic.
6631 * We temporarily dropped the "-l" switch for fsck invocations,
6632 since they collide with the flock() logic above. util-linux
6633 upstream has been changed already to avoid this conflict,
6634 and we will re-add "-l" as soon as util-linux with this
6635 change has been released.
6637 * The dependency on libattr has been removed. Since a long
6638 time, the extended attribute calls have moved to glibc, and
6639 libattr is thus unnecessary.
6641 * Virtualization detection works without privileges now. This
6642 means the systemd-detect-virt binary no longer requires
6643 CAP_SYS_PTRACE file capabilities, and our daemons can run
6644 with fewer privileges.
6646 * systemd-networkd now runs under its own "systemd-network"
6647 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
6648 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
6649 loses the ability to write to files owned by root this way.
6651 * Similarly, systemd-resolved now runs under its own
6652 "systemd-resolve" user with no capabilities remaining.
6654 * Similarly, systemd-bus-proxyd now runs under its own
6655 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
6657 * systemd-networkd gained support for setting up "veth"
6658 virtual Ethernet devices for container connectivity, as well
6659 as GRE and VTI tunnels.
6661 * systemd-networkd will no longer automatically attempt to
6662 manually load kernel modules necessary for certain tunnel
6663 transports. Instead, it is assumed the kernel loads them
6664 automatically when required. This only works correctly on
6665 very new kernels. On older kernels, please consider adding
6666 the kernel modules to /etc/modules-load.d/ as a work-around.
6668 * The resolv.conf file systemd-resolved generates has been
6669 moved to /run/systemd/resolve/. If you have a symlink from
6670 /etc/resolv.conf, it might be necessary to correct it.
6672 * Two new service settings, ProtectHome= and ProtectSystem=,
6673 have been added. When enabled, they will make the user data
6674 (such as /home) inaccessible or read-only and the system
6675 (such as /usr) read-only, for specific services. This allows
6676 very light-weight per-service sandboxing to avoid
6677 modifications of user data or system files from
6678 services. These two new switches have been enabled for all
6679 of systemd's long-running services, where appropriate.
6681 * Socket units gained new SocketUser= and SocketGroup=
6682 settings to set the owner user and group of AF_UNIX sockets
6683 and FIFOs in the file system.
6685 * Socket units gained a new RemoveOnStop= setting. If enabled,
6686 all FIFOS and sockets in the file system will be removed
6687 when the specific socket unit is stopped.
6689 * Socket units gained a new Symlinks= setting. It takes a list
6690 of symlinks to create to file system sockets or FIFOs
6691 created by the specific Unix sockets. This is useful to
6692 manage symlinks to socket nodes with the same lifecycle as
6695 * The /dev/log socket and /dev/initctl FIFO have been moved to
6696 /run, and have been replaced by symlinks. This allows
6697 connecting to these facilities even if PrivateDevices=yes is
6698 used for a service (which makes /dev/log itself unavailable,
6699 but /run is left). This also has the benefit of ensuring
6700 that /dev only contains device nodes, directories and
6701 symlinks, and nothing else.
6703 * sd-daemon gained two new calls sd_pid_notify() and
6704 sd_pid_notifyf(). They are similar to sd_notify() and
6705 sd_notifyf(), but allow overriding of the source PID of
6706 notification messages if permissions permit this. This is
6707 useful to send notify messages on behalf of a different
6708 process (for example, the parent process). The
6709 systemd-notify tool has been updated to make use of this
6710 when sending messages (so that notification messages now
6711 originate from the shell script invoking systemd-notify and
6712 not the systemd-notify process itself. This should minimize
6713 a race where systemd fails to associate notification
6714 messages to services when the originating process already
6717 * A new "on-abnormal" setting for Restart= has been added. If
6718 set, it will result in automatic restarts on all "abnormal"
6719 reasons for a process to exit, which includes unclean
6720 signals, core dumps, timeouts and watchdog timeouts, but
6721 does not include clean and unclean exit codes or clean
6722 signals. Restart=on-abnormal is an alternative for
6723 Restart=on-failure for services that shall be able to
6724 terminate and avoid restarts on certain errors, by
6725 indicating so with an unclean exit code. Restart=on-failure
6726 or Restart=on-abnormal is now the recommended setting for
6727 all long-running services.
6729 * If the InaccessibleDirectories= service setting points to a
6730 mount point (or if there are any submounts contained within
6731 it), it is now attempted to completely unmount it, to make
6732 the file systems truly unavailable for the respective
6735 * The ReadOnlyDirectories= service setting and
6736 systemd-nspawn's --read-only parameter are now recursively
6737 applied to all submounts, too.
6739 * Mount units may now be created transiently via the bus APIs.
6741 * The support for SysV and LSB init scripts has been removed
6742 from the systemd daemon itself. Instead, it is now
6743 implemented as a generator that creates native systemd units
6744 from these scripts when needed. This enables us to remove a
6745 substantial amount of legacy code from PID 1, following the
6746 fact that many distributions only ship a very small number
6747 of LSB/SysV init scripts nowadays.
6749 * Privileged Xen (dom0) domains are not considered
6750 virtualization anymore by the virtualization detection
6751 logic. After all, they generally have unrestricted access to
6752 the hardware and usually are used to manage the unprivileged
6755 * systemd-tmpfiles gained a new "C" line type, for copying
6756 files or entire directories.
6758 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
6759 lines. So far, they have been non-globbing versions of the
6760 latter, and have thus been redundant. In future, it is
6761 recommended to only use "z". "m" has hence been removed
6762 from the documentation, even though it stays supported.
6764 * A tmpfiles snippet to recreate the most basic structure in
6765 /var has been added. This is enough to create the /var/run →
6766 /run symlink and create a couple of structural
6767 directories. This allows systems to boot up with an empty or
6768 volatile /var. Of course, while with this change, the core OS
6769 now is capable with dealing with a volatile /var, not all
6770 user services are ready for it. However, we hope that sooner
6771 or later, many service daemons will be changed upstream so
6772 that they are able to automatically create their necessary
6773 directories in /var at boot, should they be missing. This is
6774 the first step to allow state-less systems that only require
6775 the vendor image for /usr to boot.
6777 * systemd-nspawn has gained a new --tmpfs= switch to mount an
6778 empty tmpfs instance to a specific directory. This is
6779 particularly useful for making use of the automatic
6780 reconstruction of /var (see above), by passing --tmpfs=/var.
6782 * Access modes specified in tmpfiles snippets may now be
6783 prefixed with "~", which indicates that they shall be masked
6784 by whether the existing file or directory is currently
6785 writable, readable or executable at all. Also, if specified,
6786 the sgid/suid/sticky bits will be masked for all
6789 * A new passive target unit "network-pre.target" has been
6790 added which is useful for services that shall run before any
6791 network is configured, for example firewall scripts.
6793 * The "floppy" group that previously owned the /dev/fd*
6794 devices is no longer used. The "disk" group is now used
6795 instead. Distributions should probably deprecate usage of
6798 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
6799 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
6800 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
6801 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
6802 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
6803 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
6806 — Berlin, 2014-06-11
6810 * A new "systemd-timesyncd" daemon has been added for
6811 synchronizing the system clock across the network. It
6812 implements an SNTP client. In contrast to NTP
6813 implementations such as chrony or the NTP reference server,
6814 this only implements a client side, and does not bother with
6815 the full NTP complexity, focusing only on querying time from
6816 one remote server and synchronizing the local clock to
6817 it. Unless you intend to serve NTP to networked clients or
6818 want to connect to local hardware clocks, this simple NTP
6819 client should be more than appropriate for most
6820 installations. The daemon runs with minimal privileges, and
6821 has been hooked up with networkd to only operate when
6822 network connectivity is available. The daemon saves the
6823 current clock to disk every time a new NTP sync has been
6824 acquired, and uses this to possibly correct the system clock
6825 early at bootup, in order to accommodate for systems that
6826 lack an RTC such as the Raspberry Pi and embedded devices,
6827 and to make sure that time monotonically progresses on these
6828 systems, even if it is not always correct. To make use of
6829 this daemon, a new system user and group "systemd-timesync"
6830 needs to be created on installation of systemd.
6832 * The queue "seqnum" interface of libudev has been disabled, as
6833 it was generally incompatible with device namespacing as
6834 sequence numbers of devices go "missing" if the devices are
6835 part of a different namespace.
6837 * "systemctl list-timers" and "systemctl list-sockets" gained
6838 a --recursive switch for showing units of these types also
6839 for all local containers, similar in style to the already
6840 supported --recursive switch for "systemctl list-units".
6842 * A new RebootArgument= setting has been added for service
6843 units, which may be used to specify a kernel reboot argument
6844 to use when triggering reboots with StartLimitAction=.
6846 * A new FailureAction= setting has been added for service
6847 units which may be used to specify an operation to trigger
6848 when a service fails. This works similarly to
6849 StartLimitAction=, but unlike it, controls what is done
6850 immediately rather than only after several attempts to
6851 restart the service in question.
6853 * hostnamed got updated to also expose the kernel name,
6854 release, and version on the bus. This is useful for
6855 executing commands like hostnamectl with the -H switch.
6856 systemd-analyze makes use of this to properly display
6857 details when running non-locally.
6859 * The bootchart tool can now show cgroup information in the
6860 graphs it generates.
6862 * The CFS CPU quota cgroup attribute is now exposed for
6863 services. The new CPUQuota= switch has been added for this
6864 which takes a percentage value. Setting this will have the
6865 result that a service may never get more CPU time than the
6866 specified percentage, even if the machine is otherwise idle.
6868 * systemd-networkd learned IPIP and SIT tunnel support.
6870 * LSB init scripts exposing a dependency on $network will now
6871 get a dependency on network-online.target rather than simply
6872 network.target. This should bring LSB handling closer to
6873 what it was on SysV systems.
6875 * A new fsck.repair= kernel option has been added to control
6876 how fsck shall deal with unclean file systems at boot.
6878 * The (.ini) configuration file parser will now silently
6879 ignore sections whose name begins with "X-". This may be
6880 used to maintain application-specific extension sections in unit
6883 * machined gained a new API to query the IP addresses of
6884 registered containers. "machinectl status" has been updated
6885 to show these addresses in its output.
6887 * A new call sd_uid_get_display() has been added to the
6888 sd-login APIs for querying the "primary" session of a
6889 user. The "primary" session of the user is elected from the
6890 user's sessions and generally a graphical session is
6891 preferred over a text one.
6893 * A minimal systemd-resolved daemon has been added. It
6894 currently simply acts as a companion to systemd-networkd and
6895 manages resolv.conf based on per-interface DNS
6896 configuration, possibly supplied via DHCP. In the long run
6897 we hope to extend this into a local DNSSEC enabled DNS and
6900 * The systemd-networkd-wait-online tool is now enabled by
6901 default. It will delay network-online.target until a network
6902 connection has been configured. The tool primarily integrates
6903 with networkd, but will also make a best effort to make sense
6904 of network configuration performed in some other way.
6906 * Two new service options StartupCPUShares= and
6907 StartupBlockIOWeight= have been added that work similarly to
6908 CPUShares= and BlockIOWeight= however only apply during
6909 system startup. This is useful to prioritize certain services
6910 differently during bootup than during normal runtime.
6912 * hostnamed has been changed to prefer the statically
6913 configured hostname in /etc/hostname (unless set to
6914 'localhost' or empty) over any dynamic one supplied by
6915 dhcp. With this change, the rules for picking the hostname
6916 match more closely the rules of other configuration settings
6917 where the local administrator's configuration in /etc always
6918 overrides any other settings.
6920 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
6921 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
6922 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
6923 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
6924 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
6925 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
6926 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
6927 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
6928 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
6929 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
6930 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
6931 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
6932 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
6933 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
6934 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
6935 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6938 — Beijing, 2014-05-28
6942 * When restoring the screen brightness at boot, stay away from
6943 the darkest setting or from the lowest 5% of the available
6944 range, depending on which is the larger value of both. This
6945 should effectively protect the user from rebooting into a
6946 black screen, should the brightness have been set to minimum
6949 * sd-login gained a new sd_machine_get_class() call to
6950 determine the class ("vm" or "container") of a machine
6951 registered with machined.
6953 * sd-login gained new calls
6954 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
6955 to query the identity of the peer of a local AF_UNIX
6956 connection. They operate similarly to their sd_pid_get_xyz()
6959 * PID 1 will now maintain a system-wide system state engine
6960 with the states "starting", "running", "degraded",
6961 "maintenance", "stopping". These states are bound to system
6962 startup, normal runtime, runtime with at least one failed
6963 service, rescue/emergency mode and system shutdown. This
6964 state is shown in the "systemctl status" output when no unit
6965 name is passed. It is useful to determine system state, in
6966 particularly when doing so for many systems or containers at
6969 * A new command "list-machines" has been added to "systemctl"
6970 that lists all local OS containers and shows their system
6971 state (see above), if systemd runs inside of them.
6973 * systemctl gained a new "-r" switch to recursively enumerate
6974 units on all local containers, when used with the
6975 "list-unit" command (which is the default one that is
6976 executed when no parameters are specified).
6978 * The GPT automatic partition discovery logic will now honour
6979 two GPT partition flags: one may be set on a partition to
6980 cause it to be mounted read-only, and the other may be set
6981 on a partition to ignore it during automatic discovery.
6983 * Two new GPT type UUIDs have been added for automatic root
6984 partition discovery, for 32-bit and 64-bit ARM. This is not
6985 particularly useful for discovering the root directory on
6986 these architectures during bare-metal boots (since UEFI is
6987 not common there), but still very useful to allow booting of
6988 ARM disk images in nspawn with the -i option.
6990 * MAC addresses of interfaces created with nspawn's
6991 --network-interface= switch will now be generated from the
6992 machine name, and thus be stable between multiple invocations
6995 * logind will now automatically remove all IPC objects owned
6996 by a user if she or he fully logs out. This makes sure that
6997 users who are logged out cannot continue to consume IPC
6998 resources. This covers SysV memory, semaphores and message
6999 queues as well as POSIX shared memory and message
7000 queues. Traditionally, SysV and POSIX IPC had no lifecycle
7001 limits. With this functionality, that is corrected. This may
7002 be turned off by using the RemoveIPC= switch of logind.conf.
7004 * The systemd-machine-id-setup and tmpfiles tools gained a
7005 --root= switch to operate on a specific root directory,
7008 * journald can now forward logged messages to the TTYs of all
7009 logged in users ("wall"). This is the default for all
7010 emergency messages now.
7012 * A new tool systemd-journal-remote has been added to stream
7013 journal log messages across the network.
7015 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
7016 controller trees are mounted into it. Note that the
7017 directories mounted beneath it are not read-only. This is a
7018 security measure and is particularly useful because glibc
7019 actually includes a search logic to pick any tmpfs it can
7020 find to implement shm_open() if /dev/shm is not available
7021 (which it might very well be in namespaced setups).
7023 * machinectl gained a new "poweroff" command to cleanly power
7024 down a local OS container.
7026 * The PrivateDevices= unit file setting will now also drop the
7027 CAP_MKNOD capability from the capability bound set, and
7028 imply DevicePolicy=closed.
7030 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
7031 comprehensively on all long-running systemd services where
7032 this is appropriate.
7034 * systemd-udevd will now run in a disassociated mount
7035 namespace. To mount directories from udev rules, make sure to
7036 pull in mount units via SYSTEMD_WANTS properties.
7038 * The kdbus support gained support for uploading policy into
7039 the kernel. sd-bus gained support for creating "monitoring"
7040 connections that can eavesdrop into all bus communication
7041 for debugging purposes.
7043 * Timestamps may now be specified in seconds since the UNIX
7044 epoch Jan 1st, 1970 by specifying "@" followed by the value
7047 * Native tcpwrap support in systemd has been removed. tcpwrap
7048 is old code, not really maintained anymore and has serious
7049 shortcomings, and better options such as firewalls
7050 exist. For setups that require tcpwrap usage, please
7051 consider invoking your socket-activated service via tcpd,
7052 like on traditional inetd.
7054 * A new system.conf configuration option
7055 DefaultTimerAccuracySec= has been added that controls the
7056 default AccuracySec= setting of .timer units.
7058 * Timer units gained a new WakeSystem= switch. If enabled,
7059 timers configured this way will cause the system to resume
7060 from system suspend (if the system supports that, which most
7063 * Timer units gained a new Persistent= switch. If enabled,
7064 timers configured this way will save to disk when they have
7065 been last triggered. This information is then used on next
7066 reboot to possible execute overdue timer events, that
7067 could not take place because the system was powered off.
7068 This enables simple anacron-like behaviour for timer units.
7070 * systemctl's "list-timers" will now also list the time a
7071 timer unit was last triggered in addition to the next time
7072 it will be triggered.
7074 * systemd-networkd will now assign predictable IPv4LL
7075 addresses to its local interfaces.
7077 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
7078 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
7079 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
7080 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
7081 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
7082 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
7083 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
7084 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
7087 — Berlin, 2014-03-25
7091 * A new unit file setting RestrictAddressFamilies= has been
7092 added to restrict which socket address families unit
7093 processes gain access to. This takes address family names
7094 like "AF_INET" or "AF_UNIX", and is useful to minimize the
7095 attack surface of services via exotic protocol stacks. This
7096 is built on seccomp system call filters.
7098 * Two new unit file settings RuntimeDirectory= and
7099 RuntimeDirectoryMode= have been added that may be used to
7100 manage a per-daemon runtime directories below /run. This is
7101 an alternative for setting up directory permissions with
7102 tmpfiles snippets, and has the advantage that the runtime
7103 directory's lifetime is bound to the daemon runtime and that
7104 the daemon starts up with an empty directory each time. This
7105 is particularly useful when writing services that drop
7106 privileges using the User= or Group= setting.
7108 * The DeviceAllow= unit setting now supports globbing for
7109 matching against device group names.
7111 * The systemd configuration file system.conf gained new
7112 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
7113 DefaultMemoryAccounting= to globally turn on/off accounting
7114 for specific resources (cgroups) for all units. These
7115 settings may still be overridden individually in each unit
7118 * systemd-gpt-auto-generator is now able to discover /srv and
7119 root partitions in addition to /home and swap partitions. It
7120 also supports LUKS-encrypted partitions now. With this in
7121 place, automatic discovery of partitions to mount following
7122 the Discoverable Partitions Specification
7123 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
7124 is now a lot more complete. This allows booting without
7125 /etc/fstab and without root= on the kernel command line on
7126 systems prepared appropriately.
7128 * systemd-nspawn gained a new --image= switch which allows
7129 booting up disk images and Linux installations on any block
7130 device that follow the Discoverable Partitions Specification
7131 (see above). This means that installations made with
7132 appropriately updated installers may now be started and
7133 deployed using container managers, completely
7134 unmodified. (We hope that libvirt-lxc will add support for
7135 this feature soon, too.)
7137 * systemd-nspawn gained a new --network-macvlan= setting to
7138 set up a private macvlan interface for the
7139 container. Similarly, systemd-networkd gained a new
7140 Kind=macvlan setting in .netdev files.
7142 * systemd-networkd now supports configuring local addresses
7145 * A new tool systemd-network-wait-online has been added to
7146 synchronously wait for network connectivity using
7149 * The sd-bus.h bus API gained a new sd_bus_track object for
7150 tracking the lifecycle of bus peers. Note that sd-bus.h is
7151 still not a public API though (unless you specify
7152 --enable-kdbus on the configure command line, which however
7153 voids your warranty and you get no API stability guarantee).
7155 * The $XDG_RUNTIME_DIR runtime directories for each user are
7156 now individual tmpfs instances, which has the benefit of
7157 introducing separate pools for each user, with individual
7158 size limits, and thus making sure that unprivileged clients
7159 can no longer negatively impact the system or other users by
7160 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
7161 RuntimeDirectorySize= has been introduced that allows
7162 controlling the default size limit for all users. It
7163 defaults to 10% of the available physical memory. This is no
7164 replacement for quotas on tmpfs though (which the kernel
7165 still does not support), as /dev/shm and /tmp are still
7166 shared resources used by both the system and unprivileged
7169 * logind will now automatically turn off automatic suspending
7170 on laptop lid close when more than one display is
7171 connected. This was previously expected to be implemented
7172 individually in desktop environments (such as GNOME),
7173 however has been added to logind now, in order to fix a
7174 boot-time race where a desktop environment might not have
7175 been started yet and thus not been able to take an inhibitor
7176 lock at the time where logind already suspends the system
7177 due to a closed lid.
7179 * logind will now wait at least 30s after each system
7180 suspend/resume cycle, and 3min after system boot before
7181 suspending the system due to a closed laptop lid. This
7182 should give USB docking stations and similar enough time to
7183 be probed and configured after system resume and boot in
7184 order to then act as suspend blocker.
7186 * systemd-run gained a new --property= setting which allows
7187 initialization of resource control properties (and others)
7188 for the created scope or service unit. Example: "systemd-run
7189 --property=BlockIOWeight=10 updatedb" may be used to run
7190 updatedb at a low block IO scheduling weight.
7192 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
7193 now also work in --scope mode.
7195 * When systemd is compiled with kdbus support, basic support
7196 for enforced policies is now in place. (Note that enabling
7197 kdbus still voids your warranty and no API compatibility
7200 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
7201 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7202 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
7203 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
7204 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
7205 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
7206 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
7207 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
7208 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
7209 Zbigniew Jędrzejewski-Szmek
7211 — Berlin, 2014-03-12
7215 * systemd will now relabel /dev after loading the SMACK policy
7216 according to SMACK rules.
7218 * A new unit file option AppArmorProfile= has been added to
7219 set the AppArmor profile for the processes of a unit.
7221 * A new condition check ConditionArchitecture= has been added
7222 to conditionalize units based on the system architecture, as
7223 reported by uname()'s "machine" field.
7225 * systemd-networkd now supports matching on the system
7226 virtualization, architecture, kernel command line, host name
7229 * logind is now a lot more aggressive when suspending the
7230 machine due to a closed laptop lid. Instead of acting only
7231 on the lid close action, it will continuously watch the lid
7232 status and act on it. This is useful for laptops where the
7233 power button is on the outside of the chassis so that it can
7234 be reached without opening the lid (such as the Lenovo
7235 Yoga). On those machines, logind will now immediately
7236 re-suspend the machine if the power button has been
7237 accidentally pressed while the laptop was suspended and in a
7238 backpack or similar.
7240 * logind will now watch SW_DOCK switches and inhibit reaction
7241 to the lid switch if it is pressed. This means that logind
7242 will not suspend the machine anymore if the lid is closed
7243 and the system is docked, if the laptop supports SW_DOCK
7244 notifications via the input layer. Note that ACPI docking
7245 stations do not generate this currently. Also note that this
7246 logic is usually not fully sufficient and Desktop
7247 Environments should take a lid switch inhibitor lock when an
7248 external display is connected, as systemd will not watch
7251 * nspawn will now make use of the devices cgroup controller by
7252 default, and only permit creation of and access to the usual
7253 API device nodes like /dev/null or /dev/random, as well as
7254 access to (but not creation of) the pty devices.
7256 * We will now ship a default .network file for
7257 systemd-networkd that automatically configures DHCP for
7258 network interfaces created by nspawn's --network-veth or
7259 --network-bridge= switches.
7261 * systemd will now understand the usual M, K, G, T suffixes
7262 according to SI conventions (i.e. to the base 1000) when
7263 referring to throughput and hardware metrics. It will stay
7264 with IEC conventions (i.e. to the base 1024) for software
7265 metrics, according to what is customary according to
7266 Wikipedia. We explicitly document which base applies for
7267 each configuration option.
7269 * The DeviceAllow= setting in unit files now supports a syntax
7270 to whitelist an entire group of devices node majors at once,
7271 based on the /proc/devices listing. For example, with the
7272 string "char-pts", it is now possible to whitelist all
7273 current and future pseudo-TTYs at once.
7275 * sd-event learned a new "post" event source. Event sources of
7276 this type are triggered by the dispatching of any event
7277 source of a type that is not "post". This is useful for
7278 implementing clean-up and check event sources that are
7279 triggered by other work being done in the program.
7281 * systemd-networkd is no longer statically enabled, but uses
7282 the usual [Install] sections so that it can be
7283 enabled/disabled using systemctl. It still is enabled by
7286 * When creating a veth interface pair with systemd-nspawn, the
7287 host side will now be prefixed with "vb-" if
7288 --network-bridge= is used, and with "ve-" if --network-veth
7289 is used. This way, it is easy to distinguish these cases on
7290 the host, for example to apply different configuration to
7291 them with systemd-networkd.
7293 * The compatibility libraries for libsystemd-journal.so,
7294 libsystem-id128.so, libsystemd-login.so and
7295 libsystemd-daemon.so do not make use of IFUNC
7296 anymore. Instead, we now build libsystemd.so multiple times
7297 under these alternative names. This means that the footprint
7298 is drastically increased, but given that these are
7299 transitional compatibility libraries, this should not matter
7300 much. This change has been made necessary to support the ARM
7301 platform for these compatibility libraries, as the ARM
7302 toolchain is not really at the same level as the toolchain
7303 for other architectures like x86 and does not support
7304 IFUNC. Please make sure to use --enable-compat-libs only
7305 during a transitional period!
7307 * The .include syntax has been deprecated and is not documented
7308 anymore. Drop-in files in .d directories should be used instead.
7310 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
7311 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7312 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
7313 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
7314 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
7315 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
7316 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
7317 Zbigniew Jędrzejewski-Szmek
7319 — Berlin, 2014-02-24
7323 * A new component "systemd-networkd" has been added that can
7324 be used to configure local network interfaces statically or
7325 via DHCP. It is capable of bringing up bridges, VLANs, and
7326 bonding. Currently, no hook-ups for interactive network
7327 configuration are provided. Use this for your initrd,
7328 container, embedded, or server setup if you need a simple,
7329 yet powerful, network configuration solution. This
7330 configuration subsystem is quite nifty, as it allows wildcard
7331 hotplug matching in interfaces. For example, with a single
7332 configuration snippet, you can configure that all Ethernet
7333 interfaces showing up are automatically added to a bridge,
7334 or similar. It supports link-sensing and more.
7336 * A new tool "systemd-socket-proxyd" has been added which can
7337 act as a bidirectional proxy for TCP sockets. This is
7338 useful for adding socket activation support to services that
7339 do not actually support socket activation, including virtual
7340 machines and the like.
7342 * Add a new tool to save/restore rfkill state on
7345 * Save/restore state of keyboard backlights in addition to
7346 display backlights on shutdown/boot.
7348 * udev learned a new SECLABEL{} construct to label device
7349 nodes with a specific security label when they appear. For
7350 now, only SECLABEL{selinux} is supported, but the syntax is
7351 prepared for additional security frameworks.
7353 * udev gained a new scheme to configure link-level attributes
7354 from files in /etc/systemd/network/*.link. These files can
7355 match against MAC address, device path, driver name and type,
7356 and will apply attributes like the naming policy, link speed,
7357 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
7358 address assignment policy (randomized, ...).
7360 * The configuration of network interface naming rules for
7361 "permanent interface names" has changed: a new NamePolicy=
7362 setting in the [Link] section of .link files determines the
7363 priority of possible naming schemes (onboard, slot, MAC,
7364 path). The default value of this setting is determined by
7365 /usr/lib/net/links/99-default.link. Old
7366 80-net-name-slot.rules udev configuration file has been
7367 removed, so local configuration overriding this file should
7368 be adapted to override 99-default.link instead.
7370 * When the User= switch is used in a unit file, also
7371 initialize $SHELL= based on the user database entry.
7373 * systemd no longer depends on libdbus. All communication is
7374 now done with sd-bus, systemd's low-level bus library
7377 * kdbus support has been added to PID 1 itself. When kdbus is
7378 enabled, this causes PID 1 to set up the system bus and
7379 enable support for a new ".busname" unit type that
7380 encapsulates bus name activation on kdbus. It works a little
7381 bit like ".socket" units, except for bus names. A new
7382 generator has been added that converts classic dbus1 service
7383 activation files automatically into native systemd .busname
7386 * sd-bus: add a light-weight vtable implementation that allows
7387 defining objects on the bus with a simple static const
7388 vtable array of its methods, signals and properties.
7390 * systemd will not generate or install static dbus
7391 introspection data anymore to /usr/share/dbus-1/interfaces,
7392 as the precise format of these files is unclear, and
7393 nothing makes use of it.
7395 * A proxy daemon is now provided to proxy clients connecting
7396 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
7397 compatibility with classic D-Bus.
7399 * A bus driver implementation has been added that supports the
7400 classic D-Bus bus driver calls on kdbus, also for
7401 compatibility purposes.
7403 * A new API "sd-event.h" has been added that implements a
7404 minimal event loop API built around epoll. It provides a
7405 couple of features that direct epoll usage is lacking:
7406 prioritization of events, scales to large numbers of timer
7407 events, per-event timer slack (accuracy), system-wide
7408 coalescing of timer events, exit handlers, watchdog
7409 supervision support using systemd's sd_notify() API, child
7412 * A new API "sd-rntl.h" has been added that provides an API
7413 around the route netlink interface of the kernel, similar in
7414 style to "sd-bus.h".
7416 * A new API "sd-dhcp-client.h" has been added that provides a
7417 small DHCPv4 client-side implementation. This is used by
7420 * There is a new kernel command line option
7421 "systemd.restore_state=0|1". When set to "0", none of the
7422 systemd tools will restore saved runtime state to hardware
7423 devices. More specifically, the rfkill and backlight states
7426 * The FsckPassNo= compatibility option in mount/service units
7427 has been removed. The fstab generator will now add the
7428 necessary dependencies automatically, and does not require
7429 PID1's support for that anymore.
7431 * journalctl gained a new switch, --list-boots, that lists
7432 recent boots with their times and boot IDs.
7434 * The various tools like systemctl, loginctl, timedatectl,
7435 busctl, systemd-run, ... have gained a new switch "-M" to
7436 connect to a specific, local OS container (as direct
7437 connection, without requiring SSH). This works on any
7438 container that is registered with machined, such as those
7439 created by libvirt-lxc or nspawn.
7441 * systemd-run and systemd-analyze also gained support for "-H"
7442 to connect to remote hosts via SSH. This is particularly
7443 useful for systemd-run because it enables queuing of jobs
7444 onto remote systems.
7446 * machinectl gained a new command "login" to open a getty
7447 login in any local container. This works with any container
7448 that is registered with machined (such as those created by
7449 libvirt-lxc or nspawn), and which runs systemd inside.
7451 * machinectl gained a new "reboot" command that may be used to
7452 trigger a reboot on a specific container that is registered
7453 with machined. This works on any container that runs an init
7454 system of some kind.
7456 * systemctl gained a new "list-timers" command to print a nice
7457 listing of installed timer units with the times they elapse
7460 * Alternative reboot() parameters may now be specified on the
7461 "systemctl reboot" command line and are passed to the
7462 reboot() system call.
7464 * systemctl gained a new --job-mode= switch to configure the
7465 mode to queue a job with. This is a more generic version of
7466 --fail, --irreversible, and --ignore-dependencies, which are
7467 still available but not advertised anymore.
7469 * /etc/systemd/system.conf gained new settings to configure
7470 various default timeouts of units, as well as the default
7471 start limit interval and burst. These may still be overridden
7474 * PID1 will now export on the bus profile data of the security
7475 policy upload process (such as the SELinux policy upload to
7478 * journald: when forwarding logs to the console, include
7479 timestamps (following the setting in
7480 /sys/module/printk/parameters/time).
7482 * OnCalendar= in timer units now understands the special
7483 strings "yearly" and "annually". (Both are equivalent)
7485 * The accuracy of timer units is now configurable with the new
7486 AccuracySec= setting. It defaults to 1min.
7488 * A new dependency type JoinsNamespaceOf= has been added that
7489 allows running two services within the same /tmp and network
7490 namespace, if PrivateNetwork= or PrivateTmp= are used.
7492 * A new command "cat" has been added to systemctl. It outputs
7493 the original unit file of a unit, and concatenates the
7494 contents of additional "drop-in" unit file snippets, so that
7495 the full configuration is shown.
7497 * systemctl now supports globbing on the various "list-xyz"
7498 commands, like "list-units" or "list-sockets", as well as on
7499 those commands which take multiple unit names.
7501 * journalctl's --unit= switch gained support for globbing.
7503 * All systemd daemons now make use of the watchdog logic so
7504 that systemd automatically notices when they hang.
7506 * If the $container_ttys environment variable is set,
7507 getty-generator will automatically spawn a getty for each
7508 listed tty. This is useful for container managers to request
7509 login gettys to be spawned on as many ttys as needed.
7511 * %h, %s, %U specifier support is not available anymore when
7512 used in unit files for PID 1. This is because NSS calls are
7513 not safe from PID 1. They stay available for --user
7514 instances of systemd, and as special case for the root user.
7516 * loginctl gained a new "--no-legend" switch to turn off output
7519 * The "sd-login.h" API gained three new calls:
7520 sd_session_is_remote(), sd_session_get_remote_user(),
7521 sd_session_get_remote_host() to query information about
7524 * The udev hardware database now also carries vendor/product
7525 information of SDIO devices.
7527 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
7528 determine whether watchdog notifications are requested by
7531 * Socket-activated per-connection services now include a
7532 short description of the connection parameters in the
7535 * tmpfiles gained a new "--boot" option. When this is not used,
7536 only lines where the command character is not suffixed with
7537 "!" are executed. When this option is specified, those
7538 options are executed too. This partitions tmpfiles
7539 directives into those that can be safely executed at any
7540 time, and those which should be run only at boot (for
7541 example, a line that creates /run/nologin).
7543 * A new API "sd-resolve.h" has been added which provides a simple
7544 asynchronous wrapper around glibc NSS host name resolution
7545 calls, such as getaddrinfo(). In contrast to glibc's
7546 getaddrinfo_a(), it does not use signals. In contrast to most
7547 other asynchronous name resolution libraries, this one does
7548 not reimplement DNS, but reuses NSS, so that alternate
7549 host name resolution systems continue to work, such as mDNS,
7550 LDAP, etc. This API is based on libasyncns, but it has been
7551 cleaned up for inclusion in systemd.
7553 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
7554 "sd-daemon.h" are no longer found in individual libraries
7555 libsystemd-journal.so, libsystemd-login.so,
7556 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
7557 merged them into a single library, libsystemd.so, which
7558 provides all symbols. The reason for this is cyclic
7559 dependencies, as these libraries tend to use each other's
7560 symbols. So far, we have managed to workaround that by linking
7561 a copy of a good part of our code into each of these
7562 libraries again and again, which, however, makes certain
7563 things hard to do, like sharing static variables. Also, it
7564 substantially increases footprint. With this change, there
7565 is only one library for the basic APIs systemd
7566 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
7567 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
7568 library as well, however are subject to the --enable-kdbus
7569 switch (see below). Note that "sd-dhcp-client.h" is not part
7570 of this library (this is because it only consumes, never
7571 provides, services of/to other APIs). To make the transition
7572 easy from the separate libraries to the unified one, we
7573 provide the --enable-compat-libs compile-time switch which
7574 will generate stub libraries that are compatible with the
7575 old ones but redirect all calls to the new one.
7577 * All of the kdbus logic and the new APIs "sd-bus.h",
7578 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
7579 and "sd-utf8.h" are compile-time optional via the
7580 "--enable-kdbus" switch, and they are not compiled in by
7581 default. To make use of kdbus, you have to explicitly enable
7582 the switch. Note however, that neither the kernel nor the
7583 userspace API for all of this is considered stable yet. We
7584 want to maintain the freedom to still change the APIs for
7585 now. By specifying this build-time switch, you acknowledge
7586 that you are aware of the instability of the current
7589 * Also, note that while kdbus is pretty much complete,
7590 it lacks one thing: proper policy support. This means you
7591 can build a fully working system with all features; however,
7592 it will be highly insecure. Policy support will be added in
7593 one of the next releases, at the same time that we will
7594 declare the APIs stable.
7596 * When the kernel command line argument "kdbus" is specified,
7597 systemd will automatically load the kdbus.ko kernel module. At
7598 this stage of development, it is only useful for testing kdbus
7599 and should not be used in production. Note: if "--enable-kdbus"
7600 is specified, and the kdbus.ko kernel module is available, and
7601 "kdbus" is added to the kernel command line, the entire system
7602 runs with kdbus instead of dbus-daemon, with the above mentioned
7603 problem of missing the system policy enforcement. Also a future
7604 version of kdbus.ko or a newer systemd will not be compatible with
7605 each other, and will unlikely be able to boot the machine if only
7606 one of them is updated.
7608 * systemctl gained a new "import-environment" command which
7609 uploads the caller's environment (or parts thereof) into the
7610 service manager so that it is inherited by services started
7611 by the manager. This is useful to upload variables like
7612 $DISPLAY into the user service manager.
7614 * A new PrivateDevices= switch has been added to service units
7615 which allows running a service with a namespaced /dev
7616 directory that does not contain any device nodes for
7617 physical devices. More specifically, it only includes devices
7618 such as /dev/null, /dev/urandom, and /dev/zero which are API
7621 * logind has been extended to support behaviour like VT
7622 switching on seats that do not support a VT. This makes
7623 multi-session available on seats that are not the first seat
7624 (seat0), and on systems where kernel support for VTs has
7625 been disabled at compile-time.
7627 * If a process holds a delay lock for system sleep or shutdown
7628 and fails to release it in time, we will now log its
7629 identity. This makes it easier to identify processes that
7630 cause slow suspends or power-offs.
7632 * When parsing /etc/crypttab, support for a new key-slot=
7633 option as supported by Debian is added. It allows indicating
7634 which LUKS slot to use on disk, speeding up key loading.
7636 * The sd_journal_sendv() API call has been checked and
7637 officially declared to be async-signal-safe so that it may
7638 be invoked from signal handlers for logging purposes.
7640 * Boot-time status output is now enabled automatically after a
7641 short timeout if boot does not progress, in order to give
7642 the user an indication what she or he is waiting for.
7644 * The boot-time output has been improved to show how much time
7645 remains until jobs expire.
7647 * The KillMode= switch in service units gained a new possible
7648 value "mixed". If set, and the unit is shut down, then the
7649 initial SIGTERM signal is sent only to the main daemon
7650 process, while the following SIGKILL signal is sent to
7651 all remaining processes of the service.
7653 * When a scope unit is registered, a new property "Controller"
7654 may be set. If set to a valid bus name, systemd will send a
7655 RequestStop() signal to this name when it would like to shut
7656 down the scope. This may be used to hook manager logic into
7657 the shutdown logic of scope units. Also, scope units may now
7658 be put in a special "abandoned" state, in which case the
7659 manager process which created them takes no further
7660 responsibilities for it.
7662 * When reading unit files, systemd will now verify
7663 the access mode of these files, and warn about certain
7664 suspicious combinations. This has been added to make it
7665 easier to track down packaging bugs where unit files are
7666 marked executable or world-writable.
7668 * systemd-nspawn gained a new "--setenv=" switch to set
7669 container-wide environment variables. The similar option in
7670 systemd-activate was renamed from "--environment=" to
7671 "--setenv=" for consistency.
7673 * systemd-nspawn has been updated to create a new kdbus domain
7674 for each container that is invoked, thus allowing each
7675 container to have its own set of system and user buses,
7676 independent of the host.
7678 * systemd-nspawn gained a new --drop-capability= switch to run
7679 the container with less capabilities than the default. Both
7680 --drop-capability= and --capability= now take the special
7681 string "all" for dropping or keeping all capabilities.
7683 * systemd-nspawn gained new switches for executing containers
7684 with specific SELinux labels set.
7686 * systemd-nspawn gained a new --quiet switch to not generate
7687 any additional output but the container's own console
7690 * systemd-nspawn gained a new --share-system switch to run a
7691 container without PID namespacing enabled.
7693 * systemd-nspawn gained a new --register= switch to control
7694 whether the container is registered with systemd-machined or
7695 not. This is useful for containers that do not run full
7696 OS images, but only specific apps.
7698 * systemd-nspawn gained a new --keep-unit which may be used
7699 when invoked as the only program from a service unit, and
7700 results in registration of the unit service itself in
7701 systemd-machined, instead of a newly opened scope unit.
7703 * systemd-nspawn gained a new --network-interface= switch for
7704 moving arbitrary interfaces to the container. The new
7705 --network-veth switch creates a virtual Ethernet connection
7706 between host and container. The new --network-bridge=
7707 switch then allows assigning the host side of this virtual
7708 Ethernet connection to a bridge device.
7710 * systemd-nspawn gained a new --personality= switch for
7711 setting the kernel personality for the container. This is
7712 useful when running a 32-bit container on a 64-bit host. A
7713 similar option Personality= is now also available for service
7716 * logind will now also track a "Desktop" identifier for each
7717 session which encodes the desktop environment of it. This is
7718 useful for desktop environments that want to identify
7719 multiple running sessions of itself easily.
7721 * A new SELinuxContext= setting for service units has been
7722 added that allows setting a specific SELinux execution
7723 context for a service.
7725 * Most systemd client tools will now honour $SYSTEMD_LESS for
7726 settings of the "less" pager. By default, these tools will
7727 override $LESS to allow certain operations to work, such as
7728 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
7729 influence this logic.
7731 * systemd's "seccomp" hook-up has been changed to make use of
7732 the libseccomp library instead of using its own
7733 implementation. This has benefits for portability among
7736 * For usage together with SystemCallFilter=, a new
7737 SystemCallErrorNumber= setting has been introduced that
7738 allows configuration of a system error number to be returned
7739 on filtered system calls, instead of immediately killing the
7740 process. Also, SystemCallArchitectures= has been added to
7741 limit access to system calls of a particular architecture
7742 (in order to turn off support for unused secondary
7743 architectures). There is also a global
7744 SystemCallArchitectures= setting in system.conf now to turn
7745 off support for non-native system calls system-wide.
7747 * systemd requires a kernel with a working name_to_handle_at(),
7748 please see the kernel config requirements in the README file.
7750 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
7751 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
7752 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
7753 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
7754 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
7755 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
7756 Elia Pinto, Florian Weimer, George McCollister, Goffredo
7757 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
7758 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
7759 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
7760 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
7761 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
7762 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
7763 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
7764 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
7765 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
7766 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
7767 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
7768 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
7769 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
7770 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
7771 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
7772 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
7773 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
7775 — Berlin, 2014-02-20
7779 * logind has gained support for facilitating privileged input
7780 and drm device access for unprivileged clients. This work is
7781 useful to allow Wayland display servers (and similar
7782 programs, such as kmscon) to run under the user's ID and
7783 access input and drm devices which are normally
7784 protected. When this is used (and the kernel is new enough)
7785 logind will "mute" IO on the file descriptors passed to
7786 Wayland as long as it is in the background and "unmute" it
7787 if it returns into the foreground. This allows secure
7788 session switching without allowing background sessions to
7789 eavesdrop on input and display data. This also introduces
7790 session switching support if VT support is turned off in the
7791 kernel, and on seats that are not seat0.
7793 * A new kernel command line option luks.options= is understood
7794 now which allows specifying LUKS options for usage for LUKS
7795 encrypted partitions specified with luks.uuid=.
7797 * tmpfiles.d(5) snippets may now use specifier expansion in
7798 path names. More specifically %m, %b, %H, %v, are now
7799 replaced by the local machine id, boot id, hostname, and
7800 kernel version number.
7802 * A new tmpfiles.d(5) command "m" has been introduced which
7803 may be used to change the owner/group/access mode of a file
7804 or directory if it exists, but do nothing if it does not.
7806 * This release removes high-level support for the
7807 MemorySoftLimit= cgroup setting. The underlying kernel
7808 cgroup attribute memory.soft_limit= is currently badly
7809 designed and likely to be removed from the kernel API in its
7810 current form, hence we should not expose it for now.
7812 * The memory.use_hierarchy cgroup attribute is now enabled for
7813 all cgroups systemd creates in the memory cgroup
7814 hierarchy. This option is likely to be come the built-in
7815 default in the kernel anyway, and the non-hierarchical mode
7816 never made much sense in the intrinsically hierarchical
7819 * A new field _SYSTEMD_SLICE= is logged along with all journal
7820 messages containing the slice a message was generated
7821 from. This is useful to allow easy per-customer filtering of
7822 logs among other things.
7824 * systemd-journald will no longer adjust the group of journal
7825 files it creates to the "systemd-journal" group. Instead we
7826 rely on the journal directory to be owned by the
7827 "systemd-journal" group, and its setgid bit set, so that the
7828 kernel file system layer will automatically enforce that
7829 journal files inherit this group assignment. The reason for
7830 this change is that we cannot allow NSS look-ups from
7831 journald which would be necessary to resolve
7832 "systemd-journal" to a numeric GID, because this might
7833 create deadlocks if NSS involves synchronous queries to
7834 other daemons (such as nscd, or sssd) which in turn are
7835 logging clients of journald and might block on it, which
7836 would then dead lock. A tmpfiles.d(5) snippet included in
7837 systemd will make sure the setgid bit and group are
7838 properly set on the journal directory if it exists on every
7839 boot. However, we recommend adjusting it manually after
7840 upgrades too (or from RPM scriptlets), so that the change is
7841 not delayed until next reboot.
7843 * Backlight and random seed files in /var/lib/ have moved into
7844 the /var/lib/systemd/ directory, in order to centralize all
7845 systemd generated files in one directory.
7847 * Boot time performance measurements (as displayed by
7848 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
7849 performance information if that's available to determine how
7850 much time BIOS and boot loader initialization required. With
7851 a sufficiently new BIOS you hence no longer need to boot
7852 with Gummiboot to get access to such information.
7854 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
7855 Cristian Rodríguez, Dave Reisner, David Herrmann, David
7856 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
7857 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
7858 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
7859 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
7860 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7862 — Berlin, 2013-10-02
7866 * The Restart= option for services now understands a new
7867 on-watchdog setting, which will restart the service
7868 automatically if the service stops sending out watchdog keep
7869 alive messages (as configured with WatchdogSec=).
7871 * The getty generator (which is responsible for bringing up a
7872 getty on configured serial consoles) will no longer only
7873 start a getty on the primary kernel console but on all
7874 others, too. This makes the order in which console= is
7875 specified on the kernel command line less important.
7877 * libsystemd-logind gained a new sd_session_get_vt() call to
7878 retrieve the VT number of a session.
7880 * If the option "tries=0" is set for an entry of /etc/crypttab
7881 its passphrase is queried indefinitely instead of any
7882 maximum number of tries.
7884 * If a service with a configure PID file terminates its PID
7885 file will now be removed automatically if it still exists
7886 afterwards. This should put an end to stale PID files.
7888 * systemd-run will now also take relative binary path names
7889 for execution and no longer insists on absolute paths.
7891 * InaccessibleDirectories= and ReadOnlyDirectories= now take
7892 paths that are optionally prefixed with "-" to indicate that
7893 it should not be considered a failure if they do not exist.
7895 * journalctl -o (and similar commands) now understands a new
7896 output mode "short-precise", it is similar to "short" but
7897 shows timestamps with usec accuracy.
7899 * The option "discard" (as known from Debian) is now
7900 synonymous to "allow-discards" in /etc/crypttab. In fact,
7901 "discard" is preferred now (since it is easier to remember
7904 * Some licensing clean-ups were made, so that more code is now
7905 LGPL-2.1 licensed than before.
7907 * A minimal tool to save/restore the display backlight
7908 brightness across reboots has been added. It will store the
7909 backlight setting as late as possible at shutdown, and
7910 restore it as early as possible during reboot.
7912 * A logic to automatically discover and enable home and swap
7913 partitions on GPT disks has been added. With this in place
7914 /etc/fstab becomes optional for many setups as systemd can
7915 discover certain partitions located on the root disk
7916 automatically. Home partitions are recognized under their
7917 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
7918 partitions are recognized under their GPT type ID
7919 0657fd6da4ab43c484e50933c84b4f4f.
7921 * systemd will no longer pass any environment from the kernel
7922 or initrd to system services. If you want to set an
7923 environment for all services, do so via the kernel command
7924 line systemd.setenv= assignment.
7926 * The systemd-sysctl tool no longer natively reads the file
7927 /etc/sysctl.conf. If desired, the file should be symlinked
7928 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
7929 legacy support by a symlink rather than built-in code, it
7930 also makes the otherwise hidden order of application of the
7931 different files visible. (Note that this partly reverts to a
7932 pre-198 application order of sysctl knobs!)
7934 * The "systemctl set-log-level" and "systemctl dump" commands
7935 have been moved to systemd-analyze.
7937 * systemd-run learned the new --remain-after-exit switch,
7938 which causes the scope unit not to be cleaned up
7939 automatically after the process terminated.
7941 * tmpfiles learned a new --exclude-prefix= switch to exclude
7942 certain paths from operation.
7944 * journald will now automatically flush all messages to disk
7945 as soon as a message at the log level CRIT, ALERT or EMERG
7948 Contributions from: Andrew Cook, Brandon Philips, Christian
7949 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
7950 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
7951 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
7952 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
7953 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
7954 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
7955 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
7956 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
7957 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
7958 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
7959 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
7960 William Giokas, Zbigniew Jędrzejewski-Szmek
7962 — Berlin, 2013-09-13
7966 * The documentation has been updated to cover the various new
7967 concepts introduced with 205.
7969 * Unit files now understand the new %v specifier which
7970 resolves to the kernel version string as returned by "uname
7973 * systemctl now supports filtering the unit list output by
7974 load state, active state and sub state, using the new
7977 * "systemctl status" will now show the results of the
7978 condition checks (like ConditionPathExists= and similar) of
7979 the last start attempts of the unit. They are also logged to
7982 * "journalctl -b" may now be used to look for boot output of a
7983 specific boot. Try "journalctl -b -1" for the previous boot,
7984 but the syntax is substantially more powerful.
7986 * "journalctl --show-cursor" has been added which prints the
7987 cursor string the last shown log line. This may then be used
7988 with the new "journalctl --after-cursor=" switch to continue
7989 browsing logs from that point on.
7991 * "journalctl --force" may now be used to force regeneration
7994 * Creation of "dead" device nodes has been moved from udev
7995 into kmod and tmpfiles. Previously, udev would read the kmod
7996 databases to pre-generate dead device nodes based on meta
7997 information contained in kernel modules, so that these would
7998 be auto-loaded on access rather then at boot. As this
7999 does not really have much to do with the exposing actual
8000 kernel devices to userspace this has always been slightly
8001 alien in the udev codebase. Following the new scheme kmod
8002 will now generate a runtime snippet for tmpfiles from the
8003 module meta information and it now is tmpfiles' job to the
8004 create the nodes. This also allows overriding access and
8005 other parameters for the nodes using the usual tmpfiles
8006 facilities. As side effect this allows us to remove the
8007 CAP_SYS_MKNOD capability bit from udevd entirely.
8009 * logind's device ACLs may now be applied to these "dead"
8010 devices nodes too, thus finally allowing managed access to
8011 devices such as /dev/snd/sequencer without loading the
8012 backing module right-away.
8014 * A new RPM macro has been added that may be used to apply
8015 tmpfiles configuration during package installation.
8017 * systemd-detect-virt and ConditionVirtualization= now can
8018 detect User-Mode-Linux machines (UML).
8020 * journald will now implicitly log the effective capabilities
8021 set of processes in the message metadata.
8023 * systemd-cryptsetup has gained support for TrueCrypt volumes.
8025 * The initrd interface has been simplified (more specifically,
8026 support for passing performance data via environment
8027 variables and fsck results via files in /run has been
8028 removed). These features were non-essential, and are
8029 nowadays available in a much nicer way by having systemd in
8030 the initrd serialize its state and have the hosts systemd
8031 deserialize it again.
8033 * The udev "keymap" data files and tools to apply keyboard
8034 specific mappings of scan to key codes, and force-release
8035 scan code lists have been entirely replaced by a udev
8036 "keyboard" builtin and a hwdb data file.
8038 * systemd will now honour the kernel's "quiet" command line
8039 argument also during late shutdown, resulting in a
8040 completely silent shutdown when used.
8042 * There's now an option to control the SO_REUSEPORT socket
8043 option in .socket units.
8045 * Instance units will now automatically get a per-template
8046 subslice of system.slice unless something else is explicitly
8047 configured. For example, instances of sshd@.service will now
8048 implicitly be placed in system-sshd.slice rather than
8049 system.slice as before.
8051 * Test coverage support may now be enabled at build time.
8053 Contributions from: Dave Reisner, Frederic Crozat, Harald
8054 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
8055 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
8056 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
8057 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
8058 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
8059 Giokas, Zbigniew Jędrzejewski-Szmek
8061 — Berlin, 2013-07-23
8065 * Two new unit types have been introduced:
8067 Scope units are very similar to service units, however, are
8068 created out of pre-existing processes — instead of PID 1
8069 forking off the processes. By using scope units it is
8070 possible for system services and applications to group their
8071 own child processes (worker processes) in a powerful way
8072 which then maybe used to organize them, or kill them
8073 together, or apply resource limits on them.
8075 Slice units may be used to partition system resources in an
8076 hierarchical fashion and then assign other units to them. By
8077 default there are now three slices: system.slice (for all
8078 system services), user.slice (for all user sessions),
8079 machine.slice (for VMs and containers).
8081 Slices and scopes have been introduced primarily in
8082 context of the work to move cgroup handling to a
8083 single-writer scheme, where only PID 1
8084 creates/removes/manages cgroups.
8086 * There's a new concept of "transient" units. In contrast to
8087 normal units these units are created via an API at runtime,
8088 not from configuration from disk. More specifically this
8089 means it is now possible to run arbitrary programs as
8090 independent services, with all execution parameters passed
8091 in via bus APIs rather than read from disk. Transient units
8092 make systemd substantially more dynamic then it ever was,
8093 and useful as a general batch manager.
8095 * logind has been updated to make use of scope and slice units
8096 for managing user sessions. As a user logs in he will get
8097 his own private slice unit, to which all sessions are added
8098 as scope units. We also added support for automatically
8099 adding an instance of user@.service for the user into the
8100 slice. Effectively logind will no longer create cgroup
8101 hierarchies on its own now, it will defer entirely to PID 1
8102 for this by means of scope, service and slice units. Since
8103 user sessions this way become entities managed by PID 1
8104 the output of "systemctl" is now a lot more comprehensive.
8106 * A new mini-daemon "systemd-machined" has been added which
8107 may be used by virtualization managers to register local
8108 VMs/containers. nspawn has been updated accordingly, and
8109 libvirt will be updated shortly. machined will collect a bit
8110 of meta information about the VMs/containers, and assign
8111 them their own scope unit (see above). The collected
8112 meta-data is then made available via the "machinectl" tool,
8113 and exposed in "ps" and similar tools. machined/machinectl
8114 is compile-time optional.
8116 * As discussed earlier, the low-level cgroup configuration
8117 options ControlGroup=, ControlGroupModify=,
8118 ControlGroupPersistent=, ControlGroupAttribute= have been
8119 removed. Please use high-level attribute settings instead as
8120 well as slice units.
8122 * A new bus call SetUnitProperties() has been added to alter
8123 various runtime parameters of a unit. This is primarily
8124 useful to alter cgroup parameters dynamically in a nice way,
8125 but will be extended later on to make more properties
8126 modifiable at runtime. systemctl gained a new set-properties
8127 command that wraps this call.
8129 * A new tool "systemd-run" has been added which can be used to
8130 run arbitrary command lines as transient services or scopes,
8131 while configuring a number of settings via the command
8132 line. This tool is currently very basic, however already
8133 very useful. We plan to extend this tool to even allow
8134 queuing of execution jobs with time triggers from the
8135 command line, similar in fashion to "at".
8137 * nspawn will now inform the user explicitly that kernels with
8138 audit enabled break containers, and suggest the user to turn
8141 * Support for detecting the IMA and AppArmor security
8142 frameworks with ConditionSecurity= has been added.
8144 * journalctl gained a new "-k" switch for showing only kernel
8145 messages, mimicking dmesg output; in addition to "--user"
8146 and "--system" switches for showing only user's own logs
8149 * systemd-delta can now show information about drop-in
8150 snippets extending unit files.
8152 * libsystemd-bus has been substantially updated but is still
8153 not available as public API.
8155 * systemd will now look for the "debug" argument on the kernel
8156 command line and enable debug logging, similar to what
8157 "systemd.log_level=debug" already did before.
8159 * "systemctl set-default", "systemctl get-default" has been
8160 added to configure the default.target symlink, which
8161 controls what to boot into by default.
8163 * "systemctl set-log-level" has been added as a convenient
8164 way to raise and lower systemd logging threshold.
8166 * "systemd-analyze plot" will now show the time the various
8167 generators needed for execution, as well as information
8168 about the unit file loading.
8170 * libsystemd-journal gained a new sd_journal_open_files() call
8171 for opening specific journal files. journactl also gained a
8172 new switch to expose this new functionality. Previously we
8173 only supported opening all files from a directory, or all
8174 files from the system, as opening individual files only is
8175 racy due to journal file rotation.
8177 * systemd gained the new DefaultEnvironment= setting in
8178 /etc/systemd/system.conf to set environment variables for
8181 * If a privileged process logs a journal message with the
8182 OBJECT_PID= field set, then journald will automatically
8183 augment this with additional OBJECT_UID=, OBJECT_GID=,
8184 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
8185 system services want to log events about specific client
8186 processes. journactl/systemctl has been updated to make use
8187 of this information if all log messages regarding a specific
8190 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
8191 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
8192 Reisner, David Coppa, David King, David Strauss, Eelco
8193 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
8194 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
8195 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
8196 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
8197 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
8198 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
8199 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
8200 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
8201 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
8202 Łukasz Stelmach, 장동준
8206 * The Python bindings gained some minimal support for the APIs
8207 exposed by libsystemd-logind.
8209 * ConditionSecurity= gained support for detecting SMACK. Since
8210 this condition already supports SELinux and AppArmor we only
8211 miss IMA for this. Patches welcome!
8213 Contributions from: Karol Lewandowski, Lennart Poettering,
8214 Zbigniew Jędrzejewski-Szmek
8218 * systemd-nspawn will now create /etc/resolv.conf if
8219 necessary, before bind-mounting the host's file onto it.
8221 * systemd-nspawn will now store meta information about a
8222 container on the container's cgroup as extended attribute
8223 fields, including the root directory.
8225 * The cgroup hierarchy has been reworked in many ways. All
8226 objects any of the components systemd creates in the cgroup
8227 tree are now suffixed. More specifically, user sessions are
8228 now placed in cgroups suffixed with ".session", users in
8229 cgroups suffixed with ".user", and nspawn containers in
8230 cgroups suffixed with ".nspawn". Furthermore, all cgroup
8231 names are now escaped in a simple scheme to avoid collision
8232 of userspace object names with kernel filenames. This work
8233 is preparation for making these objects relocatable in the
8234 cgroup tree, in order to allow easy resource partitioning of
8235 these objects without causing naming conflicts.
8237 * systemctl list-dependencies gained the new switches
8238 --plain, --reverse, --after and --before.
8240 * systemd-inhibit now shows the process name of processes that
8241 have taken an inhibitor lock.
8243 * nss-myhostname will now also resolve "localhost"
8244 implicitly. This makes /etc/hosts an optional file and
8245 nicely handles that on IPv6 ::1 maps to both "localhost" and
8248 * libsystemd-logind.so gained a new call
8249 sd_get_machine_names() to enumerate running containers and
8250 VMs (currently only supported by very new libvirt and
8251 nspawn). sd_login_monitor can now be used to watch
8252 VMs/containers coming and going.
8254 * .include is not allowed recursively anymore, and only in
8255 unit files. Usually it is better to use drop-in snippets in
8256 .d/*.conf anyway, as introduced with systemd 198.
8258 * systemd-analyze gained a new "critical-chain" command that
8259 determines the slowest chain of units run during system
8260 boot-up. It is very useful for tracking down where
8261 optimizing boot time is the most beneficial.
8263 * systemd will no longer allow manipulating service paths in
8264 the name=systemd:/system cgroup tree using ControlGroup= in
8265 units. (But is still fine with it in all other dirs.)
8267 * There's a new systemd-nspawn@.service service file that may
8268 be used to easily run nspawn containers as system
8269 services. With the container's root directory in
8270 /var/lib/container/foobar it is now sufficient to run
8271 "systemctl start systemd-nspawn@foobar.service" to boot it.
8273 * systemd-cgls gained a new parameter "--machine" to list only
8274 the processes within a certain container.
8276 * ConditionSecurity= now can check for "apparmor". We still
8277 are lacking checks for SMACK and IMA for this condition
8278 check though. Patches welcome!
8280 * A new configuration file /etc/systemd/sleep.conf has been
8281 added that may be used to configure which kernel operation
8282 systemd is supposed to execute when "suspend", "hibernate"
8283 or "hybrid-sleep" is requested. This makes the new kernel
8284 "freeze" state accessible to the user.
8286 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
8287 the passed argument if applicable.
8289 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
8290 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
8291 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
8292 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
8293 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
8294 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
8295 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
8300 * The output of 'systemctl list-jobs' got some polishing. The
8301 '--type=' argument may now be passed more than once. A new
8302 command 'systemctl list-sockets' has been added which shows
8303 a list of kernel sockets systemd is listening on with the
8304 socket units they belong to, plus the units these socket
8307 * The experimental libsystemd-bus library got substantial
8308 updates to work in conjunction with the (also experimental)
8309 kdbus kernel project. It works well enough to exchange
8310 messages with some sophistication. Note that kdbus is not
8311 ready yet, and the library is mostly an elaborate test case
8312 for now, and not installable.
8314 * systemd gained a new unit 'systemd-static-nodes.service'
8315 that generates static device nodes earlier during boot, and
8316 can run in conjunction with udev.
8318 * libsystemd-login gained a new call sd_pid_get_user_unit()
8319 to retrieve the user systemd unit a process is running
8320 in. This is useful for systems where systemd is used as
8323 * systemd-nspawn now places all containers in the new /machine
8324 top-level cgroup directory in the name=systemd
8325 hierarchy. libvirt will soon do the same, so that we get a
8326 uniform separation of /system, /user and /machine for system
8327 services, user processes and containers/virtual
8328 machines. This new cgroup hierarchy is also useful to stick
8329 stable names to specific container instances, which can be
8330 recognized later this way (this name may be controlled
8331 via systemd-nspawn's new -M switch). libsystemd-login also
8332 gained a new call sd_pid_get_machine_name() to retrieve the
8333 name of the container/VM a specific process belongs to.
8335 * bootchart can now store its data in the journal.
8337 * libsystemd-journal gained a new call
8338 sd_journal_add_conjunction() for AND expressions to the
8339 matching logic. This can be used to express more complex
8340 logical expressions.
8342 * journactl can now take multiple --unit= and --user-unit=
8345 * The cryptsetup logic now understands the "luks.key=" kernel
8346 command line switch for specifying a file to read the
8347 decryption key from. Also, if a configured key file is not
8348 found the tool will now automatically fall back to prompting
8351 * Python systemd.journal module was updated to wrap recently
8352 added functions from libsystemd-journal. The interface was
8353 changed to bring the low level interface in s.j._Reader
8354 closer to the C API, and the high level interface in
8355 s.j.Reader was updated to wrap and convert all data about
8358 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
8359 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
8360 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
8361 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
8362 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
8363 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8367 * journalctl --update-catalog now understands a new --root=
8368 option to operate on catalogs found in a different root
8371 * During shutdown after systemd has terminated all running
8372 services a final killing loop kills all remaining left-over
8373 processes. We will now print the name of these processes
8374 when we send SIGKILL to them, since this usually indicates a
8377 * If /etc/crypttab refers to password files stored on
8378 configured mount points automatic dependencies will now be
8379 generated to ensure the specific mount is established first
8380 before the key file is attempted to be read.
8382 * 'systemctl status' will now show information about the
8383 network sockets a socket unit is listening on.
8385 * 'systemctl status' will also shown information about any
8386 drop-in configuration file for units. (Drop-In configuration
8387 files in this context are files such as
8388 /etc/systemd/systemd/foobar.service.d/*.conf)
8390 * systemd-cgtop now optionally shows summed up CPU times of
8391 cgroups. Press '%' while running cgtop to switch between
8392 percentage and absolute mode. This is useful to determine
8393 which cgroups use up the most CPU time over the entire
8394 runtime of the system. systemd-cgtop has also been updated
8395 to be 'pipeable' for processing with further shell tools.
8397 * 'hostnamectl set-hostname' will now allow setting of FQDN
8400 * The formatting and parsing of time span values has been
8401 changed. The parser now understands fractional expressions
8402 such as "5.5h". The formatter will now output fractional
8403 expressions for all time spans under 1min, i.e. "5.123456s"
8404 rather than "5s 123ms 456us". For time spans under 1s
8405 millisecond values are shown, for those under 1ms
8406 microsecond values are shown. This should greatly improve
8407 all time-related output of systemd.
8409 * libsystemd-login and libsystemd-journal gained new
8410 functions for querying the poll() events mask and poll()
8411 timeout value for integration into arbitrary event
8414 * localectl gained the ability to list available X11 keymaps
8415 (models, layouts, variants, options).
8417 * 'systemd-analyze dot' gained the ability to filter for
8418 specific units via shell-style globs, to create smaller,
8419 more useful graphs. I.e. it is now possible to create simple
8420 graphs of all the dependencies between only target units, or
8421 of all units that Avahi has dependencies with.
8423 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
8424 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
8425 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
8426 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
8427 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
8428 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
8429 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
8433 * The boot-time readahead implementation for rotating media
8434 will now read the read-ahead data in multiple passes which
8435 consist of all read requests made in equidistant time
8436 intervals. This means instead of strictly reading read-ahead
8437 data in its physical order on disk we now try to find a
8438 middle ground between physical and access time order.
8440 * /etc/os-release files gained a new BUILD_ID= field for usage
8441 on operating systems that provide continuous builds of OS
8444 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
8445 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
8446 William Douglas, Zbigniew Jędrzejewski-Szmek
8450 * systemd-python gained an API exposing libsystemd-daemon.
8452 * The SMACK setup logic gained support for uploading CIPSO
8455 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
8456 ReadOnlyDirectories= and InaccessibleDirectories= has
8457 changed. The private /tmp and /var/tmp directories are now
8458 shared by all processes of a service (which means
8459 ExecStartPre= may now leave data in /tmp that ExecStart= of
8460 the same service can still access). When a service is
8461 stopped its temporary directories are immediately deleted
8462 (normal clean-up with tmpfiles is still done in addition to
8465 * By default, systemd will now set a couple of sysctl
8466 variables in the kernel: the safe sysrq options are turned
8467 on, IP route verification is turned on, and source routing
8468 disabled. The recently added hardlink and softlink
8469 protection of the kernel is turned on. These settings should
8470 be reasonably safe, and good defaults for all new systems.
8472 * The predictable network naming logic may now be turned off
8473 with a new kernel command line switch: net.ifnames=0.
8475 * A new libsystemd-bus module has been added that implements a
8476 pretty complete D-Bus client library. For details see:
8478 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
8480 * journald will now explicitly flush the journal files to disk
8481 at the latest 5min after each write. The file will then also
8482 be marked offline until the next write. This should increase
8483 reliability in case of a crash. The synchronization delay
8484 can be configured via SyncIntervalSec= in journald.conf.
8486 * There's a new remote-fs-setup.target unit that can be used
8487 to pull in specific services when at least one remote file
8488 system is to be mounted.
8490 * There are new targets timers.target and paths.target as
8491 canonical targets to pull user timer and path units in
8492 from. This complements sockets.target with a similar
8493 purpose for socket units.
8495 * libudev gained a new call udev_device_set_attribute_value()
8496 to set sysfs attributes of a device.
8498 * The udev daemon now sets the default number of worker
8499 processes executed in parallel based on the number of available
8500 CPUs instead of the amount of available RAM. This is supposed
8501 to provide a more reliable default and limit a too aggressive
8502 parallelism for setups with 1000s of devices connected.
8504 Contributions from: Auke Kok, Colin Walters, Cristian
8505 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
8506 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
8507 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
8508 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
8509 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
8510 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
8511 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
8512 Zbigniew Jędrzejewski-Szmek
8516 * Configuration of unit files may now be extended via drop-in
8517 files without having to edit/override the unit files
8518 themselves. More specifically, if the administrator wants to
8519 change one value for a service file foobar.service he can
8520 now do so by dropping in a configuration snippet into
8521 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
8522 will load all these snippets and apply them on top of the
8523 main unit configuration file, possibly extending or
8524 overriding its settings. Using these drop-in snippets is
8525 generally nicer than the two earlier options for changing
8526 unit files locally: copying the files from
8527 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
8528 them there; or creating a new file in /etc/systemd/system/
8529 that incorporates the original one via ".include". Drop-in
8530 snippets into these .d/ directories can be placed in any
8531 directory systemd looks for units in, and the usual
8532 overriding semantics between /usr/lib, /etc and /run apply
8535 * Most unit file settings which take lists of items can now be
8536 reset by assigning the empty string to them. For example,
8537 normally, settings such as Environment=FOO=BAR append a new
8538 environment variable assignment to the environment block,
8539 each time they are used. By assigning Environment= the empty
8540 string the environment block can be reset to empty. This is
8541 particularly useful with the .d/*.conf drop-in snippets
8542 mentioned above, since this adds the ability to reset list
8543 settings from vendor unit files via these drop-ins.
8545 * systemctl gained a new "list-dependencies" command for
8546 listing the dependencies of a unit recursively.
8548 * Inhibitors are now honored and listed by "systemctl
8549 suspend", "systemctl poweroff" (and similar) too, not only
8550 GNOME. These commands will also list active sessions by
8553 * Resource limits (as exposed by the various control group
8554 controllers) can now be controlled dynamically at runtime
8555 for all units. More specifically, you can now use a command
8556 like "systemctl set-cgroup-attr foobar.service cpu.shares
8557 2000" to alter the CPU shares a specific service gets. These
8558 settings are stored persistently on disk, and thus allow the
8559 administrator to easily adjust the resource usage of
8560 services with a few simple commands. This dynamic resource
8561 management logic is also available to other programs via the
8562 bus. Almost any kernel cgroup attribute and controller is
8565 * systemd-vconsole-setup will now copy all font settings to
8566 all allocated VTs, where it previously applied them only to
8569 * libsystemd-login gained the new sd_session_get_tty() API
8572 * This release drops support for a few legacy or
8573 distribution-specific LSB facility names when parsing init
8574 scripts: $x-display-manager, $mail-transfer-agent,
8575 $mail-transport-agent, $mail-transfer-agent, $smtp,
8576 $null. Also, the mail-transfer-agent.target unit backing
8577 this has been removed. Distributions which want to retain
8578 compatibility with this should carry the burden for
8579 supporting this themselves and patch support for these back
8580 in, if they really need to. Also, the facilities $syslog and
8581 $local_fs are now ignored, since systemd does not support
8582 early-boot LSB init scripts anymore, and these facilities
8583 are implied anyway for normal services. syslog.target has
8586 * There are new bus calls on PID1's Manager object for
8587 cancelling jobs, and removing snapshot units. Previously,
8588 both calls were only available on the Job and Snapshot
8591 * systemd-journal-gatewayd gained SSL support.
8593 * The various "environment" files, such as /etc/locale.conf
8594 now support continuation lines with a backslash ("\") as
8595 last character in the line, similarly in style (but different)
8596 to how this is supported in shells.
8598 * For normal user processes the _SYSTEMD_USER_UNIT= field is
8599 now implicitly appended to every log entry logged. systemctl
8600 has been updated to filter by this field when operating on a
8601 user systemd instance.
8603 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
8604 CAP_AUDIT_CONTROL capabilities to the capabilities set for
8605 the container. This makes it easier to boot unmodified
8606 Fedora systems in a container, which however still requires
8607 audit=0 to be passed on the kernel command line. Auditing in
8608 kernel and userspace is unfortunately still too broken in
8609 context of containers, hence we recommend compiling it out
8610 of the kernel or using audit=0. Hopefully this will be fixed
8611 one day for good in the kernel.
8613 * nspawn gained the new --bind= and --bind-ro= parameters to
8614 bind mount specific directories from the host into the
8617 * nspawn will now mount its own devpts file system instance
8618 into the container, in order not to leak pty devices from
8619 the host into the container.
8621 * systemd will now read the firmware boot time performance
8622 information from the EFI variables, if the used boot loader
8623 supports this, and takes it into account for boot performance
8624 analysis via "systemd-analyze". This is currently supported
8625 only in conjunction with Gummiboot, but could be supported
8626 by other boot loaders too. For details see:
8628 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
8630 * A new generator has been added that automatically mounts the
8631 EFI System Partition (ESP) to /boot, if that directory
8632 exists, is empty, and no other file system has been
8633 configured to be mounted there.
8635 * logind will now send out PrepareForSleep(false) out
8636 unconditionally, after coming back from suspend. This may be
8637 used by applications as asynchronous notification for
8638 system resume events.
8640 * "systemctl unlock-sessions" has been added, that allows
8641 unlocking the screens of all user sessions at once, similar
8642 to how "systemctl lock-sessions" already locked all users
8643 sessions. This is backed by a new D-Bus call UnlockSessions().
8645 * "loginctl seat-status" will now show the master device of a
8646 seat. (i.e. the device of a seat that needs to be around for
8647 the seat to be considered available, usually the graphics
8650 * tmpfiles gained a new "X" line type, that allows
8651 configuration of files and directories (with wildcards) that
8652 shall be excluded from automatic cleanup ("aging").
8654 * udev default rules set the device node permissions now only
8655 at "add" events, and do not change them any longer with a
8656 later "change" event.
8658 * The log messages for lid events and power/sleep keypresses
8659 now carry a message ID.
8661 * We now have a substantially larger unit test suite, but this
8662 continues to be work in progress.
8664 * udevadm hwdb gained a new --root= parameter to change the
8665 root directory to operate relative to.
8667 * logind will now issue a background sync() request to the kernel
8668 early at shutdown, so that dirty buffers are flushed to disk early
8669 instead of at the last moment, in order to optimize shutdown
8672 * A new bootctl tool has been added that is an interface for
8673 certain boot loader operations. This is currently a preview
8674 and is likely to be extended into a small mechanism daemon
8675 like timedated, localed, hostnamed, and can be used by
8676 graphical UIs to enumerate available boot options, and
8677 request boot into firmware operations.
8679 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
8680 the rest of the package. It also has been updated to work
8681 correctly in initrds.
8683 * polkit previously has been runtime optional, and is now also
8684 compile time optional via a configure switch.
8686 * systemd-analyze has been reimplemented in C. Also "systemctl
8687 dot" has moved into systemd-analyze.
8689 * "systemctl status" with no further parameters will now print
8690 the status of all active or failed units.
8692 * Operations such as "systemctl start" can now be executed
8693 with a new mode "--irreversible" which may be used to queue
8694 operations that cannot accidentally be reversed by a later
8695 job queuing. This is by default used to make shutdown
8696 requests more robust.
8698 * The Python API of systemd now gained a new module for
8699 reading journal files.
8701 * A new tool kernel-install has been added that can install
8702 kernel images according to the Boot Loader Specification:
8704 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
8706 * Boot time console output has been improved to provide
8707 animated boot time output for hanging jobs.
8709 * A new tool systemd-activate has been added which can be used
8710 to test socket activation with, directly from the command
8711 line. This should make it much easier to test and debug
8712 socket activation in daemons.
8714 * journalctl gained a new "--reverse" (or -r) option to show
8715 journal output in reverse order (i.e. newest line first).
8717 * journalctl gained a new "--pager-end" (or -e) option to jump
8718 to immediately jump to the end of the journal in the
8719 pager. This is only supported in conjunction with "less".
8721 * journalctl gained a new "--user-unit=" option, that works
8722 similarly to "--unit=" but filters for user units rather than
8725 * A number of unit files to ease adoption of systemd in
8726 initrds has been added. This moves some minimal logic from
8727 the various initrd implementations into systemd proper.
8729 * The journal files are now owned by a new group
8730 "systemd-journal", which exists specifically to allow access
8731 to the journal, and nothing else. Previously, we used the
8732 "adm" group for that, which however possibly covers more
8733 than just journal/log file access. This new group is now
8734 already used by systemd-journal-gatewayd to ensure this
8735 daemon gets access to the journal files and as little else
8736 as possible. Note that "make install" will also set FS ACLs
8737 up for /var/log/journal to give "adm" and "wheel" read
8738 access to it, in addition to "systemd-journal" which owns
8739 the journal files. We recommend that packaging scripts also
8740 add read access to "adm" + "wheel" to /var/log/journal, and
8741 all existing/future journal files. To normal users and
8742 administrators little changes, however packagers need to
8743 ensure to create the "systemd-journal" system group at
8744 package installation time.
8746 * The systemd-journal-gatewayd now runs as unprivileged user
8747 systemd-journal-gateway:systemd-journal-gateway. Packaging
8748 scripts need to create these system user/group at
8751 * timedated now exposes a new boolean property CanNTP that
8752 indicates whether a local NTP service is available or not.
8754 * systemd-detect-virt will now also detect xen PVs
8756 * The pstore file system is now mounted by default, if it is
8759 * In addition to the SELinux and IMA policies we will now also
8760 load SMACK policies at early boot.
8762 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
8763 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
8764 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
8765 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
8766 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
8767 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
8768 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
8769 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
8770 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
8771 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
8772 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
8773 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
8774 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
8775 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
8779 * Timer units now support calendar time events in addition to
8780 monotonic time events. That means you can now trigger a unit
8781 based on a calendar time specification such as "Thu,Fri
8782 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
8783 or fifth day of any month of the year 2013, given that it is
8784 a thursday or friday. This brings timer event support
8785 considerably closer to cron's capabilities. For details on
8786 the supported calendar time specification language see
8789 * udev now supports a number of different naming policies for
8790 network interfaces for predictable names, and a combination
8791 of these policies is now the default. Please see this wiki
8792 document for details:
8794 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8796 * Auke Kok's bootchart implementation has been added to the
8797 systemd tree. It is an optional component that can graph the
8798 boot in quite some detail. It is one of the best bootchart
8799 implementations around and minimal in its code and
8802 * nss-myhostname has been integrated into the systemd source
8803 tree. nss-myhostname guarantees that the local hostname
8804 always stays resolvable via NSS. It has been a weak
8805 requirement of systemd-hostnamed since a long time, and
8806 since its code is actually trivial we decided to just
8807 include it in systemd's source tree. It can be turned off
8808 with a configure switch.
8810 * The read-ahead logic is now capable of properly detecting
8811 whether a btrfs file system is on SSD or rotating media, in
8812 order to optimize the read-ahead scheme. Previously, it was
8813 only capable of detecting this on traditional file systems
8816 * In udev, additional device properties are now read from the
8817 IAB in addition to the OUI database. Also, Bluetooth company
8818 identities are attached to the devices as well.
8820 * In service files %U may be used as specifier that is
8821 replaced by the configured user name of the service.
8823 * nspawn may now be invoked without a controlling TTY. This
8824 makes it suitable for invocation as its own service. This
8825 may be used to set up a simple containerized server system
8826 using only core OS tools.
8828 * systemd and nspawn can now accept socket file descriptors
8829 when they are started for socket activation. This enables
8830 implementation of socket activated nspawn
8831 containers. i.e. think about autospawning an entire OS image
8832 when the first SSH or HTTP connection is received. We expect
8833 that similar functionality will also be added to libvirt-lxc
8836 * journalctl will now suppress ANSI color codes when
8837 presenting log data.
8839 * systemctl will no longer show control group information for
8840 a unit if the control group is empty anyway.
8842 * logind can now automatically suspend/hibernate/shutdown the
8845 * /etc/machine-info and hostnamed now also expose the chassis
8846 type of the system. This can be used to determine whether
8847 the local system is a laptop, desktop, handset or
8848 tablet. This information may either be configured by the
8849 user/vendor or is automatically determined from ACPI and DMI
8850 information if possible.
8852 * A number of polkit actions are now bound together with "imply"
8853 rules. This should simplify creating UIs because many actions
8854 will now authenticate similar ones as well.
8856 * Unit files learnt a new condition ConditionACPower= which
8857 may be used to conditionalize a unit depending on whether an
8858 AC power source is connected or not, of whether the system
8859 is running on battery power.
8861 * systemctl gained a new "is-failed" verb that may be used in
8862 shell scripts and suchlike to check whether a specific unit
8863 is in the "failed" state.
8865 * The EnvironmentFile= setting in unit files now supports file
8866 globbing, and can hence be used to easily read a number of
8867 environment files at once.
8869 * systemd will no longer detect and recognize specific
8870 distributions. All distribution-specific #ifdeffery has been
8871 removed, systemd is now fully generic and
8872 distribution-agnostic. Effectively, not too much is lost as
8873 a lot of the code is still accessible via explicit configure
8874 switches. However, support for some distribution specific
8875 legacy configuration file formats has been dropped. We
8876 recommend distributions to simply adopt the configuration
8877 files everybody else uses now and convert the old
8878 configuration from packaging scripts. Most distributions
8879 already did that. If that's not possible or desirable,
8880 distributions are welcome to forward port the specific
8881 pieces of code locally from the git history.
8883 * When logging a message about a unit systemd will now always
8884 log the unit name in the message meta data.
8886 * localectl will now also discover system locale data that is
8887 not stored in locale archives, but directly unpacked.
8889 * logind will no longer unconditionally use framebuffer
8890 devices as seat masters, i.e. as devices that are required
8891 to be existing before a seat is considered preset. Instead,
8892 it will now look for all devices that are tagged as
8893 "seat-master" in udev. By default, framebuffer devices will
8894 be marked as such, but depending on local systems, other
8895 devices might be marked as well. This may be used to
8896 integrate graphics cards using closed source drivers (such
8897 as NVidia ones) more nicely into logind. Note however, that
8898 we recommend using the open source NVidia drivers instead,
8899 and no udev rules for the closed-source drivers will be
8900 shipped from us upstream.
8902 Contributions from: Adam Williamson, Alessandro Crismani, Auke
8903 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
8904 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
8905 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
8906 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
8907 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
8908 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
8909 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
8910 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
8911 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
8912 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
8913 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
8918 * udev gained support for loading additional device properties
8919 from an indexed database that is keyed by vendor/product IDs
8920 and similar device identifiers. For the beginning this
8921 "hwdb" is populated with data from the well-known PCI and
8922 USB database, but also includes PNP, ACPI and OID data. In
8923 the longer run this indexed database shall grow into
8924 becoming the one central database for non-essential
8925 userspace device metadata. Previously, data from the PCI/USB
8926 database was only attached to select devices, since the
8927 lookup was a relatively expensive operation due to O(n) time
8928 complexity (with n being the number of entries in the
8929 database). Since this is now O(1), we decided to add in this
8930 data for all devices where this is available, by
8931 default. Note that the indexed database needs to be rebuilt
8932 when new data files are installed. To achieve this you need
8933 to update your packaging scripts to invoke "udevadm hwdb
8934 --update" after installation of hwdb data files. For
8935 RPM-based distributions we introduced the new
8936 %udev_hwdb_update macro for this purpose.
8938 * The Journal gained support for the "Message Catalog", an
8939 indexed database to link up additional information with
8940 journal entries. For further details please check:
8942 https://www.freedesktop.org/wiki/Software/systemd/catalog
8944 The indexed message catalog database also needs to be
8945 rebuilt after installation of message catalog files. Use
8946 "journalctl --update-catalog" for this. For RPM-based
8947 distributions we introduced the %journal_catalog_update
8948 macro for this purpose.
8950 * The Python Journal bindings gained support for the standard
8951 Python logging framework.
8953 * The Journal API gained new functions for checking whether
8954 the underlying file system of a journal file is capable of
8955 properly reporting file change notifications, or whether
8956 applications that want to reflect journal changes "live"
8957 need to recheck journal files continuously in appropriate
8960 * It is now possible to set the "age" field for tmpfiles
8961 entries to 0, indicating that files matching this entry
8962 shall always be removed when the directories are cleaned up.
8964 * coredumpctl gained a new "gdb" verb which invokes gdb
8965 right-away on the selected coredump.
8967 * There's now support for "hybrid sleep" on kernels that
8968 support this, in addition to "suspend" and "hibernate". Use
8969 "systemctl hybrid-sleep" to make use of this.
8971 * logind's HandleSuspendKey= setting (and related settings)
8972 now gained support for a new "lock" setting to simply
8973 request the screen lock on all local sessions, instead of
8974 actually executing a suspend or hibernation.
8976 * systemd will now mount the EFI variables file system by
8979 * Socket units now gained support for configuration of the
8980 SMACK security label.
8982 * timedatectl will now output the time of the last and next
8983 daylight saving change.
8985 * We dropped support for various legacy and distro-specific
8986 concepts, such as insserv, early-boot SysV services
8987 (i.e. those for non-standard runlevels such as 'b' or 'S')
8988 or ArchLinux /etc/rc.conf support. We recommend the
8989 distributions who still need support this to either continue
8990 to maintain the necessary patches downstream, or find a
8991 different solution. (Talk to us if you have questions!)
8993 * Various systemd components will now bypass polkit checks for
8994 root and otherwise handle properly if polkit is not found to
8995 be around. This should fix most issues for polkit-less
8996 systems. Quite frankly this should have been this way since
8997 day one. It is absolutely our intention to make systemd work
8998 fine on polkit-less systems, and we consider it a bug if
8999 something does not work as it should if polkit is not around.
9001 * For embedded systems it is now possible to build udev and
9002 systemd without blkid and/or kmod support.
9004 * "systemctl switch-root" is now capable of switching root
9005 more than once. I.e. in addition to transitions from the
9006 initrd to the host OS it is now possible to transition to
9007 further OS images from the host. This is useful to implement
9008 offline updating tools.
9010 * Various other additions have been made to the RPM macros
9011 shipped with systemd. Use %udev_rules_update() after
9012 installing new udev rules files. %_udevhwdbdir,
9013 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
9014 %_sysctldir are now available which resolve to the right
9015 directories for packages to place various data files in.
9017 * journalctl gained the new --full switch (in addition to
9018 --all, to disable ellipsation for long messages.
9020 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
9021 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
9022 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
9023 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
9024 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
9025 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
9026 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
9027 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
9028 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
9032 * journalctl gained new --since= and --until= switches to
9033 filter by time. It also now supports nice filtering for
9034 units via --unit=/-u.
9036 * Type=oneshot services may use ExecReload= and do the
9039 * The journal daemon now supports time-based rotation and
9040 vacuuming, in addition to the usual disk-space based
9043 * The journal will now index the available field values for
9044 each field name. This enables clients to show pretty drop
9045 downs of available match values when filtering. The bash
9046 completion of journalctl has been updated
9047 accordingly. journalctl gained a new switch -F to list all
9048 values a certain field takes in the journal database.
9050 * More service events are now written as structured messages
9051 to the journal, and made recognizable via message IDs.
9053 * The timedated, localed and hostnamed mini-services which
9054 previously only provided support for changing time, locale
9055 and hostname settings from graphical DEs such as GNOME now
9056 also have a minimal (but very useful) text-based client
9057 utility each. This is probably the nicest way to changing
9058 these settings from the command line now, especially since
9059 it lists available options and is fully integrated with bash
9062 * There's now a new tool "systemd-coredumpctl" to list and
9063 extract coredumps from the journal.
9065 * We now install a README each in /var/log/ and
9066 /etc/rc.d/init.d explaining where the system logs and init
9067 scripts went. This hopefully should help folks who go to
9068 that dirs and look into the otherwise now empty void and
9069 scratch their heads.
9071 * When user-services are invoked (by systemd --user) the
9072 $MANAGERPID env var is set to the PID of systemd.
9074 * SIGRTMIN+24 when sent to a --user instance will now result
9075 in immediate termination of systemd.
9077 * gatewayd received numerous feature additions such as a
9078 "follow" mode, for live syncing and filtering.
9080 * browse.html now allows filtering and showing detailed
9081 information on specific entries. Keyboard navigation and
9082 mouse screen support has been added.
9084 * gatewayd/journalctl now supports HTML5/JSON
9085 Server-Sent-Events as output.
9087 * The SysV init script compatibility logic will now
9088 heuristically determine whether a script supports the
9089 "reload" verb, and only then make this available as
9092 * "systemctl status --follow" has been removed, use "journalctl
9095 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
9096 have been removed since they are hardly useful to be
9099 * And I'd like to take the opportunity to specifically mention
9100 Zbigniew for his great contributions. Zbigniew, you rock!
9102 Contributions from: Andrew Eikum, Christian Hesse, Colin
9103 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
9104 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
9105 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
9106 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
9107 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
9108 Jędrzejewski-Szmek, Сковорода Никита Андреевич
9112 * If /etc/vconsole.conf is non-existent or empty we will no
9113 longer load any console font or key map at boot by
9114 default. Instead the kernel defaults will be left
9115 intact. This is definitely the right thing to do, as no
9116 configuration should mean no configuration, and hard-coding
9117 font names that are different on all archs is probably a bad
9118 idea. Also, the kernel default key map and font should be
9119 good enough for most cases anyway, and mostly identical to
9120 the userspace fonts/key maps we previously overloaded them
9121 with. If distributions want to continue to default to a
9122 non-kernel font or key map they should ship a default
9123 /etc/vconsole.conf with the appropriate contents.
9125 Contributions from: Colin Walters, Daniel J Walsh, Dave
9126 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
9127 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
9131 * journalctl gained a new --cursor= switch to show entries
9132 starting from the specified location in the journal.
9134 * We now enforce a size limit on journal entry fields exported
9135 with "-o json" in journalctl. Fields larger than 4K will be
9136 assigned null. This can be turned off with --all.
9138 * An (optional) journal gateway daemon is now available as
9139 "systemd-journal-gatewayd.service". This service provides
9140 access to the journal via HTTP and JSON. This functionality
9141 will be used to implement live log synchronization in both
9142 pull and push modes, but has various other users too, such
9143 as easy log access for debugging of embedded devices. Right
9144 now it is already useful to retrieve the journal via HTTP:
9146 # systemctl start systemd-journal-gatewayd.service
9147 # wget http://localhost:19531/entries
9149 This will download the journal contents in a
9150 /var/log/messages compatible format. The same as JSON:
9152 # curl -H"Accept: application/json" http://localhost:19531/entries
9154 This service is also accessible via a web browser where a
9155 single static HTML5 app is served that uses the JSON logic
9156 to enable the user to do some basic browsing of the
9157 journal. This will be extended later on. Here's an example
9158 screenshot of this app in its current state:
9160 http://0pointer.de/public/journal-gatewayd
9162 Contributions from: Kay Sievers, Lennart Poettering, Robert
9163 Milasan, Tom Gundersen
9167 * The bash completion logic is now available for journalctl
9170 * We do not mount the "cpuset" controller anymore together with
9171 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
9172 started if no parameters are assigned to it. "cpuset" hence
9173 broke code that assumed it could create "cpu" groups and
9176 * journalctl -f will now subscribe to terminal size changes,
9177 and line break accordingly.
9179 Contributions from: Dave Reisner, Kay Sievers, Lennart
9180 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
9184 * nspawn will now create a symlink /etc/localtime in the
9185 container environment, copying the host's timezone
9186 setting. Previously this has been done via a bind mount, but
9187 since symlinks cannot be bind mounted this has now been
9188 changed to create/update the appropriate symlink.
9190 * journalctl -n's line number argument is now optional, and
9191 will default to 10 if omitted.
9193 * journald will now log the maximum size the journal files may
9194 take up on disk. This is particularly useful if the default
9195 built-in logic of determining this parameter from the file
9196 system size is used. Use "systemctl status
9197 systemd-journald.service" to see this information.
9199 * The multi-seat X wrapper tool has been stripped down. As X
9200 is now capable of enumerating graphics devices via udev in a
9201 seat-aware way the wrapper is not strictly necessary
9202 anymore. A stripped down temporary stop-gap is still shipped
9203 until the upstream display managers have been updated to
9204 fully support the new X logic. Expect this wrapper to be
9205 removed entirely in one of the next releases.
9207 * HandleSleepKey= in logind.conf has been split up into
9208 HandleSuspendKey= and HandleHibernateKey=. The old setting
9209 is not available anymore. X11 and the kernel are
9210 distinguishing between these keys and we should too. This
9211 also means the inhibition lock for these keys has been split
9214 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
9215 Poettering, Lukas Nykryn, Václav Pavlín
9219 * Whenever a unit changes state we will now log this to the
9220 journal and show along the unit's own log output in
9223 * ConditionPathIsMountPoint= can now properly detect bind
9224 mount points too. (Previously, a bind mount of one file
9225 system to another place in the same file system could not be
9226 detected as mount, since they shared struct stat's st_dev
9229 * We will now mount the cgroup controllers cpu, cpuacct,
9230 cpuset and the controllers net_cls, net_prio together by
9233 * nspawn containers will now have a virtualized boot
9234 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
9235 over with a randomized ID at container initialization). This
9236 has the effect of making "journalctl -b" do the right thing
9239 * The JSON output journal serialization has been updated not
9240 to generate "endless" list objects anymore, but rather one
9241 JSON object per line. This is more in line how most JSON
9242 parsers expect JSON objects. The new output mode
9243 "json-pretty" has been added to provide similar output, but
9244 neatly aligned for readability by humans.
9246 * We dropped all explicit sync() invocations in the shutdown
9247 code. The kernel does this implicitly anyway in the kernel
9248 reboot() syscall. halt(8)'s -n option is now a compatibility
9251 * We now support virtualized reboot() in containers, as
9252 supported by newer kernels. We will fall back to exit() if
9253 CAP_SYS_REBOOT is not available to the container. Also,
9254 nspawn makes use of this now and will actually reboot the
9255 container if the containerized OS asks for that.
9257 * journalctl will only show local log output by default
9258 now. Use --merge (-m) to show remote log output, too.
9260 * libsystemd-journal gained the new sd_journal_get_usage()
9261 call to determine the current disk usage of all journal
9262 files. This is exposed in the new "journalctl --disk-usage"
9265 * journald gained a new configuration setting SplitMode= in
9266 journald.conf which may be used to control how user journals
9267 are split off. See journald.conf(5) for details.
9269 * A new condition type ConditionFileNotEmpty= has been added.
9271 * tmpfiles' "w" lines now support file globbing, to write
9272 multiple files at once.
9274 * We added Python bindings for the journal submission
9275 APIs. More Python APIs for a number of selected APIs will
9276 likely follow. Note that we intend to add native bindings
9277 only for the Python language, as we consider it common
9278 enough to deserve bindings shipped within systemd. There are
9279 various projects outside of systemd that provide bindings
9280 for languages such as PHP or Lua.
9282 * Many conditions will now resolve specifiers such as %i. In
9283 addition, PathChanged= and related directives of .path units
9284 now support specifiers as well.
9286 * There's now a new RPM macro definition for the system preset
9289 * journald will now warn if it ca not forward a message to the
9290 syslog daemon because its socket is full.
9292 * timedated will no longer write or process /etc/timezone,
9293 except on Debian. As we do not support late mounted /usr
9294 anymore /etc/localtime always being a symlink is now safe,
9295 and hence the information in /etc/timezone is not necessary
9298 * logind will now always reserve one VT for a text getty (VT6
9299 by default). Previously if more than 6 X sessions where
9300 started they took up all the VTs with auto-spawned gettys,
9301 so that no text gettys were available anymore.
9303 * udev will now automatically inform the btrfs kernel logic
9304 about btrfs RAID components showing up. This should make
9305 simple hotplug based btrfs RAID assembly work.
9307 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
9308 (but not for its children which will stay at the kernel
9309 default). This should allow setups with a lot more listening
9312 * systemd will now always pass the configured timezone to the
9313 kernel at boot. timedated will do the same when the timezone
9316 * logind's inhibition logic has been updated. By default,
9317 logind will now handle the lid switch, the power and sleep
9318 keys all the time, even in graphical sessions. If DEs want
9319 to handle these events on their own they should take the new
9320 handle-power-key, handle-sleep-key and handle-lid-switch
9321 inhibitors during their runtime. A simple way to achieve
9322 that is to invoke the DE wrapped in an invocation of:
9324 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
9326 * Access to unit operations is now checked via SELinux taking
9327 the unit file label and client process label into account.
9329 * systemd will now notify the administrator in the journal
9330 when he over-mounts a non-empty directory.
9332 * There are new specifiers that are resolved in unit files,
9333 for the host name (%H), the machine ID (%m) and the boot ID
9336 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
9337 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
9338 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
9339 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
9340 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
9341 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
9342 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
9346 * Support for reading structured kernel messages from
9347 /dev/kmsg has now been added and is enabled by default.
9349 * Support for reading kernel messages from /proc/kmsg has now
9350 been removed. If you want kernel messages in the journal
9351 make sure to run a recent kernel (>= 3.5) that supports
9352 reading structured messages from /dev/kmsg (see
9353 above). /proc/kmsg is now exclusive property of classic
9354 syslog daemons again.
9356 * The libudev API gained the new
9357 udev_device_new_from_device_id() call.
9359 * The logic for file system namespace (ReadOnlyDirectory=,
9360 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
9361 require pivot_root() anymore. This means fewer temporary
9362 directories are created below /tmp for this feature.
9364 * nspawn containers will now see and receive all submounts
9365 made on the host OS below the root file system of the
9368 * Forward Secure Sealing is now supported for Journal files,
9369 which provide cryptographical sealing of journal files so
9370 that attackers cannot alter log history anymore without this
9371 being detectable. Lennart will soon post a blog story about
9372 this explaining it in more detail.
9374 * There are two new service settings RestartPreventExitStatus=
9375 and SuccessExitStatus= which allow configuration of exit
9376 status (exit code or signal) which will be excepted from the
9377 restart logic, resp. consider successful.
9379 * journalctl gained the new --verify switch that can be used
9380 to check the integrity of the structure of journal files and
9381 (if Forward Secure Sealing is enabled) the contents of
9384 * nspawn containers will now be run with /dev/stdin, /dev/fd/
9385 and similar symlinks pre-created. This makes running shells
9386 as container init process a lot more fun.
9388 * The fstab support can now handle PARTUUID= and PARTLABEL=
9391 * A new ConditionHost= condition has been added to match
9392 against the hostname (with globs) and machine ID. This is
9393 useful for clusters where a single OS image is used to
9394 provision a large number of hosts which shall run slightly
9395 different sets of services.
9397 * Services which hit the restart limit will now be placed in a
9400 Contributions from: Bertram Poettering, Dave Reisner, Huang
9401 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
9402 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
9406 * When running in --user mode systemd will now become a
9407 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
9408 tree a lot more organized.
9410 * A new PartOf= unit dependency type has been introduced that
9411 may be used to group services in a natural way.
9413 * "systemctl enable" may now be used to enable instances of
9416 * journalctl now prints error log levels in red, and
9417 warning/notice log levels in bright white. It also supports
9418 filtering by log level now.
9420 * cgtop gained a new -n switch (similar to top), to configure
9421 the maximum number of iterations to run for. It also gained
9422 -b, to run in batch mode (accepting no input).
9424 * The suffix ".service" may now be omitted on most systemctl
9425 command lines involving service unit names.
9427 * There's a new bus call in logind to lock all sessions, as
9428 well as a loginctl verb for it "lock-sessions".
9430 * libsystemd-logind.so gained a new call sd_journal_perror()
9431 that works similar to libc perror() but logs to the journal
9432 and encodes structured information about the error number.
9434 * /etc/crypttab entries now understand the new keyfile-size=
9437 * shutdown(8) now can send a (configurable) wall message when
9438 a shutdown is cancelled.
9440 * The mount propagation mode for the root file system will now
9441 default to "shared", which is useful to make containers work
9442 nicely out-of-the-box so that they receive new mounts from
9443 the host. This can be undone locally by running "mount
9444 --make-rprivate /" if needed.
9446 * The prefdm.service file has been removed. Distributions
9447 should maintain this unit downstream if they intend to keep
9448 it around. However, we recommend writing normal unit files
9449 for display managers instead.
9451 * Since systemd is a crucial part of the OS we will now
9452 default to a number of compiler switches that improve
9453 security (hardening) such as read-only relocations, stack
9454 protection, and suchlike.
9456 * The TimeoutSec= setting for services is now split into
9457 TimeoutStartSec= and TimeoutStopSec= to allow configuration
9458 of individual time outs for the start and the stop phase of
9461 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
9462 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
9463 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
9464 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
9465 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
9466 Gundersen, Zbigniew Jędrzejewski-Szmek
9470 * The journal and id128 C APIs are now fully documented as man
9473 * Extra safety checks have been added when transitioning from
9474 the initial RAM disk to the main system to avoid accidental
9477 * /etc/crypttab entries now understand the new keyfile-offset=
9480 * systemctl -t can now be used to filter by unit load state.
9482 * The journal C API gained the new sd_journal_wait() call to
9483 make writing synchronous journal clients easier.
9485 * journalctl gained the new -D switch to show journals from a
9488 * journalctl now displays a special marker between log
9489 messages of two different boots.
9491 * The journal is now explicitly flushed to /var via a service
9492 systemd-journal-flush.service, rather than implicitly simply
9493 by seeing /var/log/journal to be writable.
9495 * journalctl (and the journal C APIs) can now match for much
9496 more complex expressions, with alternatives and
9499 * When transitioning from the initial RAM disk to the main
9500 system we will now kill all processes in a killing spree to
9501 ensure no processes stay around by accident.
9503 * Three new specifiers may be used in unit files: %u, %h, %s
9504 resolve to the user name, user home directory resp. user
9505 shell. This is useful for running systemd user instances.
9507 * We now automatically rotate journal files if their data
9508 object hash table gets a fill level > 75%. We also size the
9509 hash table based on the configured maximum file size. This
9510 together should lower hash collisions drastically and thus
9511 speed things up a bit.
9513 * journalctl gained the new "--header" switch to introspect
9514 header data of journal files.
9516 * A new setting SystemCallFilters= has been added to services
9517 which may be used to apply blacklists or whitelists to
9518 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
9520 * nspawn gained a new --link-journal= switch (and quicker: -j)
9521 to link the container journal with the host. This makes it
9522 very easy to centralize log viewing on the host for all
9523 guests while still keeping the journal files separated.
9525 * Many bugfixes and optimizations
9527 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
9528 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
9529 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
9534 * Several tools now understand kernel command line arguments,
9535 which are only read when run in an initial RAM disk. They
9536 usually follow closely their normal counterparts, but are
9539 * There's a new tool to analyze the readahead files that are
9540 automatically generated at boot. Use:
9542 /usr/lib/systemd/systemd-readahead analyze /.readahead
9544 * We now provide an early debug shell on tty9 if this enabled. Use:
9546 systemctl enable debug-shell.service
9548 * All plymouth related units have been moved into the Plymouth
9549 package. Please make sure to upgrade your Plymouth version
9552 * systemd-tmpfiles now supports getting passed the basename of
9553 a configuration file only, in which case it will look for it
9554 in all appropriate directories automatically.
9556 * udevadm info now takes a /dev or /sys path as argument, and
9557 does the right thing. Example:
9559 udevadm info /dev/sda
9560 udevadm info /sys/class/block/sda
9562 * systemctl now prints a warning if a unit is stopped but a
9563 unit that might trigger it continues to run. Example: a
9564 service is stopped but the socket that activates it is left
9567 * "systemctl status" will now mention if the log output was
9568 shortened due to rotation since a service has been started.
9570 * The journal API now exposes functions to determine the
9571 "cutoff" times due to rotation.
9573 * journald now understands SIGUSR1 and SIGUSR2 for triggering
9574 immediately flushing of runtime logs to /var if possible,
9575 resp. for triggering immediate rotation of the journal
9578 * It is now considered an error if a service is attempted to
9579 be stopped that is not loaded.
9581 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
9583 * systemd-analyze now supports Python 3
9585 * tmpfiles now supports cleaning up directories via aging
9586 where the first level dirs are always kept around but
9587 directories beneath it automatically aged. This is enabled
9588 by prefixing the age field with '~'.
9590 * Seat objects now expose CanGraphical, CanTTY properties
9591 which is required to deal with very fast bootups where the
9592 display manager might be running before the graphics drivers
9593 completed initialization.
9595 * Seat objects now expose a State property.
9597 * We now include RPM macros for service enabling/disabling
9598 based on the preset logic. We recommend RPM based
9599 distributions to make use of these macros if possible. This
9600 makes it simpler to reuse RPM spec files across
9603 * We now make sure that the collected systemd unit name is
9604 always valid when services log to the journal via
9607 * There's a new man page kernel-command-line(7) detailing all
9608 command line options we understand.
9610 * The fstab generator may now be disabled at boot by passing
9611 fstab=0 on the kernel command line.
9613 * A new kernel command line option modules-load= is now understood
9614 to load a specific kernel module statically, early at boot.
9616 * Unit names specified on the systemctl command line are now
9617 automatically escaped as needed. Also, if file system or
9618 device paths are specified they are automatically turned
9619 into the appropriate mount or device unit names. Example:
9621 systemctl status /home
9622 systemctl status /dev/sda
9624 * The SysVConsole= configuration option has been removed from
9625 system.conf parsing.
9627 * The SysV search path is no longer exported on the D-Bus
9630 * The Names= option has been removed from unit file parsing.
9632 * There's a new man page bootup(7) detailing the boot process.
9634 * Every unit and every generator we ship with systemd now
9635 comes with full documentation. The self-explanatory boot is
9638 * A couple of services gained "systemd-" prefixes in their
9639 name if they wrap systemd code, rather than only external
9640 code. Among them fsck@.service which is now
9641 systemd-fsck@.service.
9643 * The HaveWatchdog property has been removed from the D-Bus
9646 * systemd.confirm_spawn= on the kernel command line should now
9649 * There's a new man page crypttab(5) which details all options
9650 we actually understand.
9652 * systemd-nspawn gained a new --capability= switch to pass
9653 additional capabilities to the container.
9655 * timedated will now read known NTP implementation unit names
9656 from /usr/lib/systemd/ntp-units.d/*.list,
9657 systemd-timedated-ntp.target has been removed.
9659 * journalctl gained a new switch "-b" that lists log data of
9660 the current boot only.
9662 * The notify socket is in the abstract namespace again, in
9663 order to support daemons which chroot() at start-up.
9665 * There is a new Storage= configuration option for journald
9666 which allows configuration of where log data should go. This
9667 also provides a way to disable journal logging entirely, so
9668 that data collected is only forwarded to the console, the
9669 kernel log buffer or another syslog implementation.
9671 * Many bugfixes and optimizations
9673 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
9674 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
9675 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
9676 Shawn Landden, Tom Gundersen
9680 * "systemctl help <unit>" now shows the man page if one is
9683 * Several new man pages have been added.
9685 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
9686 MaxLevelConsole= can now be specified in
9687 journald.conf. These options allow reducing the amount of
9688 data stored on disk or forwarded by the log level.
9690 * TimerSlackNSec= can now be specified in system.conf for
9691 PID1. This allows system-wide power savings.
9693 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
9694 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
9699 * logind is now capable of (optionally) handling power and
9700 sleep keys as well as the lid switch.
9702 * journalctl now understands the syntax "journalctl
9703 /usr/bin/avahi-daemon" to get all log output of a specific
9706 * CapabilityBoundingSet= in system.conf now also influences
9707 the capability bound set of usermode helpers of the kernel.
9709 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
9710 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
9711 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
9712 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
9716 * Note that we skipped 139 releases here in order to set the
9717 new version to something that is greater than both udev's
9718 and systemd's most recent version number.
9720 * udev: all udev sources are merged into the systemd source tree now.
9721 All future udev development will happen in the systemd tree. It
9722 is still fully supported to use the udev daemon and tools without
9723 systemd running, like in initramfs or other init systems. Building
9724 udev though, will require the *build* of the systemd tree, but
9725 udev can be properly *run* without systemd.
9727 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
9728 should be used to create dead device nodes as workarounds for broken
9731 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
9732 no longer supported. udev_monitor_new_from_netlink() needs to be
9733 used to subscribe to events.
9735 * udev: when udevd is started by systemd, processes which are left
9736 behind by forking them off of udev rules, are unconditionally cleaned
9737 up and killed now after the event handling has finished. Services or
9738 daemons must be started as systemd services. Services can be
9739 pulled-in by udev to get started, but they can no longer be directly
9740 forked by udev rules.
9742 * udev: the daemon binary is called systemd-udevd now and installed
9743 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
9744 to adapt to that, create symlink, or rename the binary after building
9747 * libudev no longer provides these symbols:
9748 udev_monitor_from_socket()
9749 udev_queue_get_failed_list_entry()
9750 udev_get_{dev,sys,run}_path()
9751 The versions number was bumped and symbol versioning introduced.
9753 * systemd-loginctl and systemd-journalctl have been renamed
9754 to loginctl and journalctl to match systemctl.
9756 * The config files: /etc/systemd/systemd-logind.conf and
9757 /etc/systemd/systemd-journald.conf have been renamed to
9758 logind.conf and journald.conf. Package updates should rename
9759 the files to the new names on upgrade.
9761 * For almost all files the license is now LGPL2.1+, changed
9762 from the previous GPL2.0+. Exceptions are some minor stuff
9763 of udev (which will be changed to LGPL2.1 eventually, too),
9764 and the MIT licensed sd-daemon.[ch] library that is suitable
9765 to be used as drop-in files.
9767 * systemd and logind now handle system sleep states, in
9768 particular suspending and hibernating.
9770 * logind now implements a sleep/shutdown/idle inhibiting logic
9771 suitable for a variety of uses. Soonishly Lennart will blog
9772 about this in more detail.
9774 * var-run.mount and var-lock.mount are no longer provided
9775 (which previously bind mounted these directories to their new
9776 places). Distributions which have not converted these
9777 directories to symlinks should consider stealing these files
9778 from git history and add them downstream.
9780 * We introduced the Documentation= field for units and added
9781 this to all our shipped units. This is useful to make it
9782 easier to explore the boot and the purpose of the various
9785 * All smaller setup units (such as
9786 systemd-vconsole-setup.service) now detect properly if they
9787 are run in a container and are skipped when
9788 appropriate. This guarantees an entirely noise-free boot in
9789 Linux container environments such as systemd-nspawn.
9791 * A framework for implementing offline system updates is now
9792 integrated, for details see:
9793 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
9795 * A new service type Type=idle is available now which helps us
9796 avoiding ugly interleaving of getty output and boot status
9799 * There's now a system-wide CapabilityBoundingSet= option to
9800 globally reduce the set of capabilities for the
9801 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
9802 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
9803 even CAP_NET_ADMIN system-wide for secure systems.
9805 * There are now system-wide DefaultLimitXXX= options to
9806 globally change the defaults of the various resource limits
9807 for all units started by PID 1.
9809 * Harald Hoyer's systemd test suite has been integrated into
9810 systemd which allows easy testing of systemd builds in qemu
9811 and nspawn. (This is really awesome! Ask us for details!)
9813 * The fstab parser is now implemented as generator, not inside
9816 * systemctl will now warn you if .mount units generated from
9817 /etc/fstab are out of date due to changes in fstab that
9818 have not been read by systemd yet.
9820 * systemd is now suitable for usage in initrds. Dracut has
9821 already been updated to make use of this. With this in place
9822 initrds get a slight bit faster but primarily are much
9823 easier to introspect and debug since "systemctl status" in
9824 the host system can be used to introspect initrd services,
9825 and the journal from the initrd is kept around too.
9827 * systemd-delta has been added, a tool to explore differences
9828 between user/admin configuration and vendor defaults.
9830 * PrivateTmp= now affects both /tmp and /var/tmp.
9832 * Boot time status messages are now much prettier and feature
9833 proper english language. Booting up systemd has never been
9836 * Read-ahead pack files now include the inode number of all
9837 files to pre-cache. When the inode changes the pre-caching
9838 is not attempted. This should be nicer to deal with updated
9839 packages which might result in changes of read-ahead
9842 * We now temporaritly lower the kernel's read_ahead_kb variable
9843 when collecting read-ahead data to ensure the kernel's
9844 built-in read-ahead does not add noise to our measurements
9845 of necessary blocks to pre-cache.
9847 * There's now RequiresMountsFor= to add automatic dependencies
9848 for all mounts necessary for a specific file system path.
9850 * MountAuto= and SwapAuto= have been removed from
9851 system.conf. Mounting file systems at boot has to take place
9854 * nspawn now learned a new switch --uuid= to set the machine
9855 ID on the command line.
9857 * nspawn now learned the -b switch to automatically search
9860 * vt102 is now the default TERM for serial TTYs, upgraded from
9863 * systemd-logind now works on VT-less systems.
9865 * The build tree has been reorganized. The individual
9866 components now have directories of their own.
9868 * A new condition type ConditionPathIsReadWrite= is now available.
9870 * nspawn learned the new -C switch to create cgroups for the
9871 container in other hierarchies.
9873 * We now have support for hardware watchdogs, configurable in
9876 * The scheduled shutdown logic now has a public API.
9878 * We now mount /tmp as tmpfs by default, but this can be
9879 masked and /etc/fstab can override it.
9881 * Since udisks does not make use of /media anymore we are not
9882 mounting a tmpfs on it anymore.
9884 * journalctl gained a new --local switch to only interleave
9885 locally generated journal files.
9887 * We can now load the IMA policy at boot automatically.
9889 * The GTK tools have been split off into a systemd-ui.
9891 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
9892 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
9893 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
9894 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
9895 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
9896 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
9897 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
9898 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
9899 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
9904 * This is mostly a bugfix release
9906 * Support optional initialization of the machine ID from the
9907 KVM or container configured UUID.
9909 * Support immediate reboots with "systemctl reboot -ff"
9911 * Show /etc/os-release data in systemd-analyze output
9913 * Many bugfixes for the journal, including endianness fixes and
9914 ensuring that disk space enforcement works
9916 * sd-login.h is C++ compatible again
9918 * Extend the /etc/os-release format on request of the Debian
9921 * We now refuse non-UTF8 strings used in various configuration
9922 and unit files. This is done to ensure we do not pass invalid
9923 data over D-Bus or expose it elsewhere.
9925 * Register Mimo USB Screens as suitable for automatic seat
9928 * Read SELinux client context from journal clients in a race
9931 * Reorder configuration file lookup order. /etc now always
9932 overrides /run in order to allow the administrator to always
9933 and unconditionally override vendor-supplied or
9934 automatically generated data.
9936 * The various user visible bits of the journal now have man
9937 pages. We still lack man pages for the journal API calls
9940 * We now ship all man pages in HTML format again in the
9943 Contributions from: Dave Reisner, Dirk Eibach, Frederic
9944 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
9945 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
9950 * This is mostly a bugfix release
9952 * systems lacking /etc/os-release are no longer supported.
9954 * Various functionality updates to libsystemd-login.so
9956 * Track class of PAM logins to distinguish greeters from
9959 Contributions from: Kay Sievers, Lennart Poettering, Michael
9964 * This is an important bugfix release for v41.
9966 * Building man pages is now optional which should be useful
9967 for those building systemd from git but unwilling to install
9970 * Watchdog support for supervising services is now usable. In
9971 a future release support for hardware watchdogs
9972 (i.e. /dev/watchdog) will be added building on this.
9974 * Service start rate limiting is now configurable and can be
9975 turned off per service. When a start rate limit is hit a
9976 reboot can automatically be triggered.
9978 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
9980 Contributions from: Benjamin Franzke, Bill Nottingham,
9981 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
9982 Schmidt, Michał Górny, Piotr Drąg
9986 * The systemd binary is installed /usr/lib/systemd/systemd now;
9987 An existing /sbin/init symlink needs to be adapted with the
9990 * The code that loads kernel modules has been ported to invoke
9991 libkmod directly, instead of modprobe. This means we do not
9992 support systems with module-init-tools anymore.
9994 * Watchdog support is now already useful, but still not
9997 * A new kernel command line option systemd.setenv= is
9998 understood to set system wide environment variables
9999 dynamically at boot.
10001 * We now limit the set of capabilities of systemd-journald.
10003 * We now set SIGPIPE to ignore by default, since it only is
10004 useful in shell pipelines, and has little use in general
10005 code. This can be disabled with IgnoreSIPIPE=no in unit
10008 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
10009 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
10014 * This is mostly a bugfix release
10016 * We now expose the reason why a service failed in the
10017 "Result" D-Bus property.
10019 * Rudimentary service watchdog support (will be completed over
10020 the next few releases.)
10022 * When systemd forks off in order execute some service we will
10023 now immediately changes its argv[0] to reflect which process
10024 it will execute. This is useful to minimize the time window
10025 with a generic argv[0], which makes bootcharts more useful
10027 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
10028 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
10029 Mike Kazantsev, Ray Strode
10033 * This is mostly a test release, but incorporates many
10036 * New systemd-cgtop tool to show control groups by their
10039 * Linking against libacl for ACLs is optional again. If
10040 disabled, support tracking device access for active logins
10041 goes becomes unavailable, and so does access to the user
10042 journals by the respective users.
10044 * If a group "adm" exists, journal files are automatically
10045 owned by them, thus allow members of this group full access
10046 to the system journal as well as all user journals.
10048 * The journal now stores the SELinux context of the logging
10049 client for all entries.
10051 * Add C++ inclusion guards to all public headers
10053 * New output mode "cat" in the journal to print only text
10054 messages, without any meta data like date or time.
10056 * Include tiny X server wrapper as a temporary stop-gap to
10057 teach XOrg udev display enumeration. This is used by display
10058 managers such as gdm, and will go away as soon as XOrg
10059 learned native udev hotplugging for display devices.
10061 * Add new systemd-cat tool for executing arbitrary programs
10062 with STDERR/STDOUT connected to the journal. Can also act as
10063 BSD logger replacement, and does so by default.
10065 * Optionally store all locally generated coredumps in the
10066 journal along with meta data.
10068 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
10069 writing short strings to files (for usage for /sys), and for
10070 creating symlinks, character and block device nodes.
10072 * New unit file option ControlGroupPersistent= to make cgroups
10073 persistent, following the mechanisms outlined in
10074 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
10076 * Support multiple local RTCs in a sane way
10078 * No longer monopolize IO when replaying readahead data on
10079 rotating disks, since we might starve non-file-system IO to
10080 death, since fanotify() will not see accesses done by blkid,
10083 * Do not show kernel threads in systemd-cgls anymore, unless
10084 requested with new -k switch.
10086 Contributions from: Dan Horák, Kay Sievers, Lennart
10087 Poettering, Michal Schmidt
10091 * This is mostly a test release, but incorporates many
10094 * The git repository moved to:
10095 git://anongit.freedesktop.org/systemd/systemd
10096 ssh://git.freedesktop.org/git/systemd/systemd
10098 * First release with the journal
10099 http://0pointer.de/blog/projects/the-journal.html
10101 * The journal replaces both systemd-kmsg-syslogd and
10102 systemd-stdout-bridge.
10104 * New sd_pid_get_unit() API call in libsystemd-logind
10106 * Many systemadm clean-ups
10108 * Introduce remote-fs-pre.target which is ordered before all
10109 remote mounts and may be used to start services before all
10112 * Added Mageia support
10114 * Add bash completion for systemd-loginctl
10116 * Actively monitor PID file creation for daemons which exit in
10117 the parent process before having finished writing the PID
10118 file in the daemon process. Daemons which do this need to be
10119 fixed (i.e. PID file creation must have finished before the
10120 parent exits), but we now react a bit more gracefully to them.
10122 * Add colourful boot output, mimicking the well-known output
10123 of existing distributions.
10125 * New option PassCredentials= for socket units, for
10126 compatibility with a recent kernel ABI breakage.
10128 * /etc/rc.local is now hooked in via a generator binary, and
10129 thus will no longer act as synchronization point during
10132 * systemctl list-unit-files now supports --root=.
10134 * systemd-tmpfiles now understands two new commands: z, Z for
10135 relabelling files according to the SELinux database. This is
10136 useful to apply SELinux labels to specific files in /sys,
10137 among other things.
10139 * Output of SysV services is now forwarded to both the console
10140 and the journal by default, not only just the console.
10142 * New man pages for all APIs from libsystemd-login.
10144 * The build tree got reorganized and the build system is a
10145 lot more modular allowing embedded setups to specifically
10146 select the components of systemd they are interested in.
10148 * Support for Linux systems lacking the kernel VT subsystem is
10151 * configure's --with-rootdir= got renamed to
10152 --with-rootprefix= to follow the naming used by udev and
10155 * Unless specified otherwise we will now install to /usr instead
10156 of /usr/local by default.
10158 * Processes with '@' in argv[0][0] are now excluded from the
10159 final shut-down killing spree, following the logic explained
10161 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
10163 * All processes remaining in a service cgroup when we enter
10164 the START or START_PRE states are now killed with
10165 SIGKILL. That means it is no longer possible to spawn
10166 background processes from ExecStart= lines (which was never
10167 supported anyway, and bad style).
10169 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
10170 reloading of units together.
10172 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
10173 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
10174 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
10175 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
10176 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
projects / platform / upstream / systemd.git / blob