1 systemd System and Service Manager
3 CHANGES WITH 242 in spe:
5 * In .link files, MACAddressPolicy=persistent (the default) is changed
6 to cover more devices. For devices like bridges, tun, tap, bond, and
7 similar interfaces that do not have other identifying information,
8 the interface name is used as the basis for persistent seed for MAC
9 and IPv4LL addresses. The way that devices that were handled
10 previously is not changed, and this change is about covering more
11 devices then previously by the "persistent" policy.
13 MACAddressPolicy=random may be used to force randomized MACs and
14 IPv4LL addresses for a device if desired.
16 Hint: the log output from udev (at debug level) was enhanced to
17 clarify what policy is followed and which attributes are used.
18 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
19 may be used to view this.
21 * The .device units generated by systemd-fstab-generator and other
22 generators do not automatically pull in the corresponding .mount unit
23 as a Wants= dependency. This means that simply plugging in the device
24 will not cause the mount unit to be started automatically. But please
25 note that the mount unit may be started for other reasons, in
26 particular if it is part of local-fs.target, and any unit which
27 (transitively) depends on local-fs.target is started.
29 * networkctl list/status/lldp now accept globbing wildcards for network
30 interface names to match against all existing interfaces.
32 * The $PIDFILE environment variable is set to point the absolute path
33 configured with PIDFile= for processes of that service.
35 * The fallback DNS server list was augmented with Cloudflare public DNS
36 servers. Use `-Ddns-servers=` to set a different fallback.
38 * A new special target usb-gadget.target will be started automatically
39 when a USB Device Controller is detected (which means that the system
42 * A new unit setting CPUQuotaPeriodSec= assigns the time period
43 relatively to which the CPU time quota specified by CPUQuota= is
46 * A new unit setting ProtectHostname= may be used to prevent services
47 from modifying hostname information (even if they otherwise would
48 have privileges to do so).
50 * A new unit setting NetworkNamespacePath= may be used to specify a
51 namespace for service or socket units through a path referring to a
52 Linux network namespace pseudo-file.
54 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
55 have an effect on .socket units: when used the listening socket is
56 created within the configured network namespace instead of the host
59 * ExecStart= command lines in unit files may now be prefixed with ':'
60 in which case environment variable substitution is
61 disabled. (Supported for the other ExecXYZ= settings, too.)
63 * systemd-networkd recognizes a new operation state 'enslaved', used
64 (instead of 'degraded' or 'carrier') for interfaces which form a
65 bridge, bond, or similar, and an new 'degraded-carrier' operational
66 state used for the bond or bridge master interface when one of the
67 enslaved devices is not operational.
69 * .network files learnt the new IgnoreCarrierLoss= option for leaving
70 networks configured even if the carrier is lost.
72 * The RequiredForOnline= setting in .network files may now specify a
73 minimum operational state required for the interface to be considered
74 "online" by systemd-networkd-wait-online. Related to this
75 systemd-networkd-wait-online gained a new option --operational-state=
76 to configure the same, and its --interface= option was updated to
77 optionally also take an operational state specific for an interface.
79 * systemd-networkd now implements L2TP tunnels.
81 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
82 may be used to cause autonomous and onlink prefixes received in IPv6
83 Router Advertisements to be ignored.
85 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
86 file settings may be used to tweak bridge behaviour.
88 * The new TripleSampling= option in .network files may be used to
89 configure CAN triple sampling.
91 * A new .netdev setting PrivateKeyFile= may be used to point to private
92 key for a WireGuard interface.
94 * /etc/crypttab now supports the same-cpu-crypt and
95 submit-from-crypt-cpus options to tweak encryption work scheduling
98 * systemd-tmpfiles will now take a BSD file lock before operating on a
99 contents of directory. This may be used to temporarily exclude
100 directories from aging by taking the same lock (useful for example
101 when extracting a tarball into /tmp or /var/tmp as a privileged user,
102 which might create files with really old timestamps, which
103 nevertheless should not be deleted). For further details, see:
105 https://systemd.io/TEMPORARY_DIRECTORIES
107 * systemd-tmpfiles' h line type gained support for the
108 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
109 controlling project quota inheritance.
111 * sd-boot and bootctl now implement support for an Extended Boot Loader
112 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
113 addition to the ESP partition mounted to /efi or /boot/efi.
114 Configuration file fragments, kernels, initrds and other EFI images
115 to boot will be loaded from both the ESP and XBOOTLDR partitions.
116 The XBOOTLDR partition was previously described by the Boot Loader
117 Specification, but implementation was missing in sd-boot. Support for
118 this concept allows using the sd-boot boot loader in more
119 conservative scenarios where the boot loader itself is placed in the
120 ESP but the kernels to boot (and their metadata) in a separate
123 * A system may now be booted with systemd.volatile=overlay on the
124 kernel command line, which causes the root file system to be set up
125 an overlayfs mount combining the root-only root directory with a
126 writable tmpfs. In this setup, the underlying root device is not
127 modified, and any changes are lost at reboot.
129 * Similar, systemd-nspawn can now boot containers with a volatile
130 overlayfs root with the new --volatile=overlay switch.
132 * systemd-nspawn can now consume OCI runtime bundles using a new
133 --oci-bundle= option. This implementation is fully usable, with most
134 features in the specification implemented, but since this a lot of
135 new code and functionality, this feature should most likely not
136 be used in production yet.
138 * systemd-nspawn now supports various options described by the OCI
139 runtime specification on the command-line and in .nspawn files:
140 --inaccessible=/Inaccessible= may be used to mask parts of the file
141 system tree, --console=/--pipe may be used to configure how standard
142 input, output, and error are set up.
144 * busctl learned the `emit` verb to generate D-Bus signals.
146 * systemd-analyze cat-config may be used to gather and display
147 configuration spread over multiple files, for example system and user
148 presets, tmpfiles.d, sysusers.d, udev rules, etc.
150 * systemd-analyze calendar now takes an optional new parameter
151 --iterations= which may be used to show a maximum number of iterations
152 the specified expression will elapse next.
154 * The sd-bus C API gained support for naming method parameters in the
157 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
158 the reboot() system call expects.
160 * journalctl learnt a new --cursor-file= option that points to a file
161 from which a cursor should be loaded in the beginning and to which
162 the updated cursor should be stored at the end.
164 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
165 detected by systemd-detect-virt (and may also be used in
166 ConditionVirtualization=).
168 * The behaviour of systemd-logind may now be modified with environment
169 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
170 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
171 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
172 skip the relevant operation completely (when set to false), or to
173 create a flag file in /run/systemd (when set to true), instead of
174 actually commencing the real operation when requested. The presence
175 of /run/systemd/reboot-to-firmware-setup,
176 /run/systemd/reboot-to-boot-loader-menu, and
177 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
178 boot loader implementations to replace some steps logind performs
179 during reboot with their own operations.
181 * systemctl can be used to request a reboot into the boot loader menu
182 or a specific boot loader entry with the new --boot-load-menu= and
183 --boot-loader-entry= options to a reboot command. (This requires a
184 boot loader that supports this, for example sd-boot.)
186 * kernel-install will no longer unconditionally create the output
187 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
188 snippets, but will do only if the machine-specific parent directory
189 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
190 to create this parent directory during sd-boot installation.
192 This makes it easier to use kernel-install with plugins which support
193 a different layout of the bootloader partitions (for example grub2).
195 * During package installation (with `ninja install`), we would create
196 symlinks for systemd-networkd.service, systemd-networkd.socket,
197 systemd-resolved.service, remote-cryptsetup.target, remote-fs.target,
198 systemd-networkd-wait-online.service, and systemd-timesyncd.service
199 in /etc, as if `systemctl enable` was called for those units, to make
200 the system usable immediately after installation. Now this is not
201 done anymore, and instead calling `systemctl preset-all` is
202 recommended after the first installation of systemd.
206 * The default locale can now be configured at compile time. Otherwise,
207 a suitable default will be selected automatically (one of C.UTF-8,
210 * The version string shown by systemd and other tools now includes the
211 git commit hash when built from git. An override may be specified
212 during compilation, which is intended to be used by distributions to
213 include the package release information.
215 * systemd-cat can now filter standard input and standard error streams
216 for different syslog priorities using the new --stderr-priority=
219 * systemd-journald and systemd-journal-remote reject entries which
220 contain too many fields (CVE-2018-16865) and set limits on the
221 process' command line length (CVE-2018-16864).
223 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
226 * A new network device NamePolicy "keep" is implemented for link files,
227 and used by default in 99-default.link (the fallback configuration
228 provided by systemd). With this policy, if the network device name
229 was already set by userspace, the device will not be renamed again.
230 This matches the naming scheme that was implemented before
231 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
232 is also enabled by default, even if not specified. Effectively, this
233 means that if naming-scheme >= 240 is specified, network devices will
234 be renamed according to the configuration, even if they have been
235 renamed already, if "keep" is not specified as the naming policy in
236 the .link file. The 99-default.link file provided by systemd includes
237 "keep" for backwards compatibility, but it is recommended for user
238 installed .link files to *not* include it.
240 The "kernel" policy, which keeps kernel names declared to be
241 "persistent", now works again as documented.
243 * kernel-install script now optionally takes the paths to one or more
244 initrd files, and passes them to all plugins.
246 * The mincore() system call has been dropped from the @system-service
247 system call filter group, as it is pretty exotic and may potentially
248 used for side-channel attacks.
250 * -fPIE is dropped from compiler and linker options. Please specify
251 -Db_pie=true option to meson to build position-independent
252 executables. Note that the meson option is supported since meson-0.49.
254 * The fs.protected_regular and fs.protected_fifos sysctls, which were
255 added in Linux 4.19 to make some data spoofing attacks harder, are
256 now enabled by default. While this will hopefully improve the
257 security of most installations, it is technically a backwards
258 incompatible change; to disable these sysctls again, place the
259 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
261 fs.protected_regular = 0
262 fs.protected_fifos = 0
264 Note that the similar hardlink and symlink protection has been
265 enabled since v199, and may be disabled likewise.
267 * The files read from the EnvironmentFile= setting in unit files now
268 parse backslashes inside quotes literally, matching the behaviour of
271 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
272 now automatically become NOPs when run in a chroot() environment.
274 * The tmpfiles.d/ "C" line type will now copy directory trees not only
275 when the destination is so far missing, but also if it already exists
276 as a directory and is empty. This is useful to cater for systems
277 where directory trees are put together from multiple separate mount
278 points but otherwise empty.
280 * A new function sd_bus_close_unref() (and the associated
281 sd_bus_close_unrefp()) has been added to libsystemd, that combines
282 sd_bus_close() and sd_bus_unref() in one.
284 * udevadm control learnt a new option for --ping for testing whether a
285 systemd-udevd instance is running and reacting.
287 * udevadm trigger learnt a new option for --wait-daemon for waiting
288 systemd-udevd daemon to be initialized.
290 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
291 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
292 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
293 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
294 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
295 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
296 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
297 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
298 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
299 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
300 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
301 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
302 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
303 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
304 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
305 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
306 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
312 * NoNewPrivileges=yes has been set for all long-running services
313 implemented by systemd. Previously, this was problematic due to
314 SELinux (as this would also prohibit the transition from PID1's label
315 to the service's label). This restriction has since been lifted, but
316 an SELinux policy update is required.
317 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
319 * DynamicUser=yes is dropped from systemd-networkd.service,
320 systemd-resolved.service and systemd-timesyncd.service, which was
321 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
322 and since v236 for systemd-timesyncd.service. The users and groups
323 systemd-network, systemd-resolve and systemd-timesync are created
324 by systemd-sysusers again. Distributors or system administrators
325 may need to create these users and groups if they not exist (or need
326 to re-enable DynamicUser= for those units) while upgrading systemd.
327 Also, the clock file for systemd-timesyncd may need to move from
328 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
330 * When unit files are loaded from disk, previously systemd would
331 sometimes (depending on the unit loading order) load units from the
332 target path of symlinks in .wants/ or .requires/ directories of other
333 units. This meant that unit could be loaded from different paths
334 depending on whether the unit was requested explicitly or as a
335 dependency of another unit, not honouring the priority of directories
336 in search path. It also meant that it was possible to successfully
337 load and start units which are not found in the unit search path, as
338 long as they were requested as a dependency and linked to from
339 .wants/ or .requires/. The target paths of those symlinks are not
340 used for loading units anymore and the unit file must be found in
343 * A new service type has been added: Type=exec. It's very similar to
344 Type=simple but ensures the service manager will wait for both fork()
345 and execve() of the main service binary to complete before proceeding
346 with follow-up units. This is primarily useful so that the manager
347 propagates any errors in the preparation phase of service execution
348 back to the job that requested the unit to be started. For example,
349 consider a service that has ExecStart= set to a file system binary
350 that doesn't exist. With Type=simple starting the unit would be
351 considered instantly successful, as only fork() has to complete
352 successfully and the manager does not wait for execve(), and hence
353 its failure is seen "too late". With the new Type=exec service type
354 starting the unit will fail, as the manager will wait for the
355 execve() and notice its failure, which is then propagated back to the
358 NOTE: with the next release 241 of systemd we intend to change the
359 systemd-run tool to default to Type=exec for transient services
360 started by it. This should be mostly safe, but in specific corner
361 cases might result in problems, as the systemd-run tool will then
362 block on NSS calls (such as user name look-ups due to User=) done
363 between the fork() and execve(), which under specific circumstances
364 might cause problems. It is recommended to specify "-p Type=simple"
365 explicitly in the few cases where this applies. For regular,
366 non-transient services (i.e. those defined with unit files on disk)
367 we will continue to default to Type=simple.
369 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
370 userspace processes is set to 1024 (soft) and 4096
371 (hard). Previously, systemd passed this on unmodified to all
372 processes it forked off. With this systemd release the hard limit
373 systemd passes on is increased to 512K, overriding the kernel's
374 defaults and substantially increasing the number of simultaneous file
375 descriptors unprivileged userspace processes can allocate. Note that
376 the soft limit remains at 1024 for compatibility reasons: the
377 traditional UNIX select() call cannot deal with file descriptors >=
378 1024 and increasing the soft limit globally might thus result in
379 programs unexpectedly allocating a high file descriptor and thus
380 failing abnormally when attempting to use it with select() (of
381 course, programs shouldn't use select() anymore, and prefer
382 poll()/epoll, but the call unfortunately remains undeservedly popular
383 at this time). This change reflects the fact that file descriptor
384 handling in the Linux kernel has been optimized in more recent
385 kernels and allocating large numbers of them should be much cheaper
386 both in memory and in performance than it used to be. Programs that
387 want to take benefit of the increased limit have to "opt-in" into
388 high file descriptors explicitly by raising their soft limit. Of
389 course, when they do that they must acknowledge that they cannot use
390 select() anymore (and neither can any shared library they use — or
391 any shared library used by any shared library they use and so on).
392 Which default hard limit is most appropriate is of course hard to
393 decide. However, given reports that ~300K file descriptors are used
394 in real-life applications we believe 512K is sufficiently high as new
395 default for now. Note that there are also reports that using very
396 high hard limits (e.g. 1G) is problematic: some software allocates
397 large arrays with one element for each potential file descriptor
398 (Java, …) — a high hard limit thus triggers excessively large memory
399 allocations in these applications. Hopefully, the new default of 512K
400 is a good middle ground: higher than what real-life applications
401 currently need, and low enough for avoid triggering excessively large
402 allocations in problematic software. (And yes, somebody should fix
405 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
406 to the highest possible values, as separate accounting of file
407 descriptors is no longer necessary, as memcg tracks them correctly as
408 part of the memory accounting anyway. Thus, from the four limits on
409 file descriptors currently enforced (fs.file-max, fs.nr_open,
410 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
411 and keep only the latter two. A set of build-time options
412 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
413 has been added to revert this change in behaviour, which might be
414 an option for systems that turn off memcg in the kernel.
416 * When no /etc/locale.conf file exists (and hence no locale settings
417 are in place), systemd will now use the "C.UTF-8" locale by default,
418 and set LANG= to it. This locale is supported by various
419 distributions including Fedora, with clear indications that upstream
420 glibc is going to make it available too. This locale enables UTF-8
421 mode by default, which appears appropriate for 2018.
423 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
424 default. This effectively switches the RFC3704 Reverse Path filtering
425 from Strict mode to Loose mode. This is more appropriate for hosts
426 that have multiple links with routes to the same networks (e.g.
427 a client with a Wi-Fi and Ethernet both connected to the internet).
429 Consult the kernel documentation for details on this sysctl:
430 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
432 * CPUAccounting=yes no longer enables the CPU controller when using
433 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
434 statistics are now provided independently from the CPU controller.
436 * Support for disabling a particular cgroup controller within a sub-tree
437 has been added through the DisableControllers= directive.
439 * cgroup_no_v1=all on the kernel command line now also implies
440 using the unified cgroup hierarchy, unless one explicitly passes
441 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
443 * The new "MemoryMin=" unit file property may now be used to set the
444 memory usage protection limit of processes invoked by the unit. This
445 controls the cgroup v2 memory.min attribute. Similarly, the new
446 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
447 cgroup v2 io.latency cgroup property for configuring per-service I/O
450 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
451 to the cgroup v1 "devices" cgroup controller.
453 * systemd-escape now is able to combine --unescape with --template. It
454 also learnt a new option --instance for extracting and unescaping the
455 instance part of a unit name.
457 * sd-bus now provides the sd_bus_message_readv() which is similar to
458 sd_bus_message_read() but takes a va_list object. The pair
459 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
460 has been added for configuring the default method call timeout to
461 use. sd_bus_error_move() may be used to efficiently move the contents
462 from one sd_bus_error structure to another, invalidating the
463 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
464 be used to control whether a bus connection object is automatically
465 flushed when an sd-event loop is exited.
467 * When processing classic BSD syslog log messages, journald will now
468 save the original time-stamp string supplied in the new
469 SYSLOG_TIMESTAMP= journal field. This permits consumers to
470 reconstruct the original BSD syslog message more correctly.
472 * StandardOutput=/StandardError= in service files gained support for
473 new "append:…" parameters, for connecting STDOUT/STDERR of a service
474 to a file, and appending to it.
476 * The signal to use as last step of killing of unit processes is now
477 configurable. Previously it was hard-coded to SIGKILL, which may now
478 be overridden with the new KillSignal= setting. Note that this is the
479 signal used when regular termination (i.e. SIGTERM) does not suffice.
480 Similarly, the signal used when aborting a program in case of a
481 watchdog timeout may now be configured too (WatchdogSignal=).
483 * The XDG_SESSION_DESKTOP environment variable may now be configured in
484 the pam_systemd argument line, using the new desktop= switch. This is
485 useful to initialize it properly from a display manager without
486 having to touch C code.
488 * Most configuration options that previously accepted percentage values
489 now also accept permille values with the '‰' suffix (instead of '%').
491 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
494 * systemd-resolved's configuration file resolved.conf gained a new
495 option ReadEtcHosts= which may be used to turn off processing and
496 honoring /etc/hosts entries.
498 * The "--wait" switch may now be passed to "systemctl
499 is-system-running", in which case the tool will synchronously wait
500 until the system finished start-up.
502 * hostnamed gained a new bus call to determine the DMI product UUID.
504 * On x86-64 systemd will now prefer using the RDRAND processor
505 instruction over /dev/urandom whenever it requires randomness that
506 neither has to be crypto-grade nor should be reproducible. This
507 should substantially reduce the amount of entropy systemd requests
508 from the kernel during initialization on such systems, though not
509 reduce it to zero. (Why not zero? systemd still needs to allocate
510 UUIDs and such uniquely, which require high-quality randomness.)
512 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
513 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
514 for forcing the "Other Information" bit in IPv6 RA messages. The
515 bonding logic gained four new options AdActorSystemPriority=,
516 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
517 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
518 shuffling of flows. The tunnel logic gained a new
519 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
520 Deployment. The policy rule logic gained four new options IPProtocol=,
521 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
522 support for the MulticastToUnicast= option. networkd also gained
523 support for configuring static IPv4 ARP or IPv6 neighbor entries.
525 * .preset files (as read by 'systemctl preset') may now be used to
526 instantiate services.
528 * /etc/crypttab now understands the sector-size= option to configure
529 the sector size for an encrypted partition.
531 * Key material for encrypted disks may now be placed on a formatted
532 medium, and referenced from /etc/crypttab by the UUID of the file
533 system, followed by "=" suffixed by the path to the key file.
535 * The "collect" udev component has been removed without replacement, as
536 it is neither used nor maintained.
538 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
539 LogsDirectory=, ConfigurationDirectory= settings are used in a
540 service the executed processes will now receive a set of environment
541 variables containing the full paths of these directories.
542 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
543 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
544 are used. Note that these options may be used multiple times per
545 service in which case the resulting paths will be concatenated and
548 * Predictable interface naming has been extended to cover InfiniBand
549 NICs. They will be exposed with an "ib" prefix.
551 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
552 which case the respective line failing is ignored.
554 * .link files may now be used to configure the equivalent to the
555 "ethtool advertise" commands.
557 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
558 alternative to libudev.h. Previously, the latter was just an internal
559 wrapper around the former, but now these two APIs are exposed
562 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
563 which calculates an app-specific boot ID similar to how
564 sd_id128_get_machine_app_specific() generates an app-specific machine
567 * A new tool systemd-id128 has been added that can be used to determine
568 and generate various 128bit IDs.
570 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
573 * systemd-hibernate-resume-generator will now honor the "noresume"
574 kernel command line option, in which case it will bypass resuming
575 from any hibernated image.
577 * The systemd-sleep.conf configuration file gained new options
578 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
579 AllowHybridSleep= for prohibiting specific sleep modes even if the
582 * portablectl is now officially supported and has thus moved to
585 * bootctl learnt the two new commands "set-default" and "set-oneshot"
586 for setting the default boot loader item to boot to (either
587 persistently or only for the next boot). This is currently only
588 compatible with sd-boot, but may be implemented on other boot loaders
589 too, that follow the boot loader interface. The updated interface is
592 https://systemd.io/BOOT_LOADER_INTERFACE
594 * A new kernel command line option systemd.early_core_pattern= is now
595 understood which may be used to influence the core_pattern PID 1
596 installs during early boot.
598 * busctl learnt two new options -j and --json= for outputting method
599 call replies, properties and monitoring output in JSON.
601 * journalctl's JSON output now supports simple ANSI coloring as well as
602 a new "json-seq" mode for generating RFC7464 output.
604 * Unit files now support the %g/%G specifiers that resolve to the UNIX
605 group/GID of the service manager runs as, similar to the existing
606 %u/%U specifiers that resolve to the UNIX user/UID.
608 * systemd-logind learnt a new global configuration option
609 UserStopDelaySec= that may be set in logind.conf. It specifies how
610 long the systemd --user instance shall remain started after a user
611 logs out. This is useful to speed up repetitive re-connections of the
612 same user, as it means the user's service manager doesn't have to be
613 stopped/restarted on each iteration, but can be reused between
614 subsequent options. This setting defaults to 10s. systemd-logind also
615 exports two new properties on its Manager D-Bus objects indicating
616 whether the system's lid is currently closed, and whether the system
619 * systemd gained support for a generic boot counting logic, which
620 generically permits automatic reverting to older boot loader entries
621 if newer updated ones don't work. The boot loader side is implemented
622 in sd-boot, but is kept open for other boot loaders too. For details
625 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
627 * The SuccessAction=/FailureAction= unit file settings now learnt two
628 new parameters: "exit" and "exit-force", which result in immediate
629 exiting of the service manager, and are only useful in systemd --user
630 and container environments.
632 * Unit files gained support for a pair of options
633 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
634 exit status to use as service manager exit status when
635 SuccessAction=/FailureAction= is set to exit or exit-force.
637 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
638 options may now be used to configure the log rate limiting applied by
639 journald per-service.
641 * systemd-analyze gained a new verb "timespan" for parsing and
642 normalizing time span values (i.e. strings like "5min 7s 8us").
644 * systemd-analyze also gained a new verb "security" for analyzing the
645 security and sand-boxing settings of services in order to determine an
646 "exposure level" for them, indicating whether a service would benefit
647 from more sand-boxing options turned on for them.
649 * "systemd-analyze syscall-filter" will now also show system calls
650 supported by the local kernel but not included in any of the defined
653 * .nspawn files now understand the Ephemeral= setting, matching the
654 --ephemeral command line switch.
656 * sd-event gained the new APIs sd_event_source_get_floating() and
657 sd_event_source_set_floating() for controlling whether a specific
658 event source is "floating", i.e. destroyed along with the even loop
661 * Unit objects on D-Bus gained a new "Refs" property that lists all
662 clients that currently have a reference on the unit (to ensure it is
665 * The JoinControllers= option in system.conf is no longer supported, as
666 it didn't work correctly, is hard to support properly, is legacy (as
667 the concept only exists on cgroup v1) and apparently wasn't used.
669 * Journal messages that are generated whenever a unit enters the failed
670 state are now tagged with a unique MESSAGE_ID. Similarly, messages
671 generated whenever a service process exits are now made recognizable,
672 too. A taged message is also emitted whenever a unit enters the
673 "dead" state on success.
675 * systemd-run gained a new switch --working-directory= for configuring
676 the working directory of the service to start. A shortcut -d is
677 equivalent, setting the working directory of the service to the
678 current working directory of the invoking program. The new --shell
679 (or just -S) option has been added for invoking the $SHELL of the
680 caller as a service, and implies --pty --same-dir --wait --collect
681 --service-type=exec. Or in other words, "systemd-run -S" is now the
682 quickest way to quickly get an interactive in a fully clean and
683 well-defined system service context.
685 * machinectl gained a new verb "import-fs" for importing an OS tree
686 from a directory. Moreover, when a directory or tarball is imported
687 and single top-level directory found with the OS itself below the OS
688 tree is automatically mangled and moved one level up.
690 * systemd-importd will no longer set up an implicit btrfs loop-back
691 file system on /var/lib/machines. If one is already set up, it will
694 * A new generator "systemd-run-generator" has been added. It will
695 synthesize a unit from one or more program command lines included in
696 the kernel command line. This is very useful in container managers
699 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
701 This will run "systemd-nspawn" on an image, invoke the specified
702 command line and immediately shut down the container again, returning
703 the command line's exit code.
705 * The block device locking logic is now documented:
707 https://systemd.io/BLOCK_DEVICE_LOCKING
709 * loginctl and machinectl now optionally output the various tables in
710 JSON using the --output= switch. It is our intention to add similar
711 support to systemctl and all other commands.
713 * udevadm's query and trigger verb now optionally take a .device unit
716 * systemd-udevd's network naming logic now understands a new
717 net.naming-scheme= kernel command line switch, which may be used to
718 pick a specific version of the naming scheme. This helps stabilizing
719 interface names even as systemd/udev are updated and the naming logic
722 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
723 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
724 initialize one to all 0xFF.
726 * After loading the SELinux policy systemd will now recursively relabel
727 all files and directories listed in
728 /run/systemd/relabel-extra.d/*.relabel (which should be simple
729 newline separated lists of paths) in addition to the ones it already
730 implicitly relabels in /run, /dev and /sys. After the relabelling is
731 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
732 removed. This is useful to permit initrds (i.e. code running before
733 the SELinux policy is in effect) to generate files in the host
734 filesystem safely and ensure that the correct label is applied during
735 the transition to the host OS.
737 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
738 mknod() handling in user namespaces. Previously mknod() would always
739 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
740 but device nodes generated that way cannot be opened, and attempts to
741 open them result in EPERM. This breaks the "graceful fallback" logic
742 in systemd's PrivateDevices= sand-boxing option. This option is
743 implemented defensively, so that when systemd detects it runs in a
744 restricted environment (such as a user namespace, or an environment
745 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
746 where device nodes cannot be created the effect of PrivateDevices= is
747 bypassed (following the logic that 2nd-level sand-boxing is not
748 essential if the system systemd runs in is itself already sand-boxed
749 as a whole). This logic breaks with 4.18 in container managers where
750 user namespacing is used: suddenly PrivateDevices= succeeds setting
751 up a private /dev/ file system containing devices nodes — but when
752 these are opened they don't work.
754 At this point is is recommended that container managers utilizing
755 user namespaces that intend to run systemd in the payload explicitly
756 block mknod() with seccomp or similar, so that the graceful fallback
759 We are very sorry for the breakage and the requirement to change
760 container configurations for newer kernels. It's purely caused by an
761 incompatible kernel change. The relevant kernel developers have been
762 notified about this userspace breakage quickly, but they chose to
765 * PermissionsStartOnly= setting is deprecated (but is still supported
766 for backwards compatibility). The same functionality is provided by
767 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
770 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
773 * The naming scheme for network devices was changed to always rename
774 devices, even if they were already renamed by userspace. The "kernel"
775 policy was changed to only apply as a fallback, if no other naming
778 * The requirements to build systemd is bumped to meson-0.46 and
781 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
782 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
783 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
784 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
785 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
786 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
787 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
788 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
789 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
790 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
791 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
792 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
793 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
794 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
795 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
796 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
797 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
798 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
799 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
800 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
801 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
802 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
803 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
804 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
805 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
806 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
807 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
808 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
809 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
810 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
811 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
812 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
813 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
814 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
815 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
816 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
817 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
818 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
819 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
820 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
821 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
822 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
823 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
824 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
825 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
831 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
832 builtin will name network interfaces differently than in previous
833 versions for virtual network interfaces created with SR-IOV and NPAR
834 and for devices where the PCI network controller device does not have
835 a slot number associated.
837 SR-IOV virtual devices are now named based on the name of the parent
838 interface, with a suffix of "v<N>", where <N> is the virtual device
839 number. Previously those virtual devices were named as if completely
842 The ninth and later NPAR virtual devices will be named following the
843 scheme used for the first eight NPAR partitions. Previously those
844 devices were not renamed and the kernel default (eth<n>) was used.
846 "net_id" will also generate names for PCI devices where the PCI
847 network controller device does not have an associated slot number
848 itself, but one of its parents does. Previously those devices were
849 not renamed and the kernel default (eth<n>) was used.
851 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
852 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
853 the unit. So, it is expected that the default behavior of
854 systemd-logind is not changed. However, if distribution packagers or
855 administrators disabled or modified IPAddressDeny= setting by a
856 drop-in config file, then it may be necessary to update the file to
857 re-enable AF_INET and AF_INET6 to support network user name services,
860 * When the RestrictNamespaces= unit property is specified multiple
861 times, then the specified types are merged now. Previously, only the
862 last assignment was used. So, if distribution packagers or
863 administrators modified the setting by a drop-in config file, then it
864 may be necessary to update the file.
866 * When OnFailure= is used in combination with Restart= on a service
867 unit, then the specified units will no longer be triggered on
868 failures that result in restarting. Previously, the specified units
869 would be activated each time the unit failed, even when the unit was
870 going to be restarted automatically. This behaviour contradicted the
871 documentation. With this release the code is adjusted to match the
874 * systemd-tmpfiles will now print a notice whenever it encounters
875 tmpfiles.d/ lines referencing the /var/run/ directory. It will
876 recommend reworking them to use the /run/ directory instead (for
877 which /var/run/ is simply a symlinked compatibility alias). This way
878 systemd-tmpfiles can properly detect line conflicts and merge lines
879 referencing the same file by two paths, without having to access
882 * systemctl disable/unmask/preset/preset-all cannot be used with
883 --runtime. Previously this was allowed, but resulted in unintuitive
884 behaviour that wasn't useful. systemctl disable/unmask will now undo
885 both runtime and persistent enablement/masking, i.e. it will remove
886 any relevant symlinks both in /run and /etc.
888 * Note that all long-running system services shipped with systemd will
889 now default to a system call whitelist (rather than a blacklist, as
890 before). In particular, systemd-udevd will now enforce one too. For
891 most cases this should be safe, however downstream distributions
892 which disabled sandboxing of systemd-udevd (specifically the
893 MountFlags= setting), might want to disable this security feature
894 too, as the default whitelisting will prohibit all mount, swap,
895 reboot and clock changing operations from udev rules.
897 * sd-boot acquired new loader configuration settings to optionally turn
898 off Windows and MacOS boot partition discovery as well as
899 reboot-into-firmware menu items. It is also able to pick a better
900 screen resolution for HiDPI systems, and now provides loader
901 configuration settings to change the resolution explicitly.
903 * systemd-resolved now supports DNS-over-TLS. It's still
904 turned off by default, use DNSOverTLS=opportunistic to turn it on in
905 resolved.conf. We intend to make this the default as soon as couple
906 of additional techniques for optimizing the initial latency caused by
907 establishing a TLS/TCP connection are implemented.
909 * systemd-resolved.service and systemd-networkd.service now set
910 DynamicUser=yes. The users systemd-resolve and systemd-network are
911 not created by systemd-sysusers anymore.
913 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
914 that embedd a network facing module into any process using getpwuid()
915 or related call: the dynamic allocation of the user ID for
916 systemd-resolved.service means the service manager has to check NSS
917 if the user name is already taken when forking off the service. Since
918 the user in the common case won't be defined in /etc/passwd the
919 lookup is likely to trigger nss-ldap which in turn might use NSS to
920 ask systemd-resolved for hostname lookups. This will hence result in
921 a deadlock: a user name lookup in order to start
922 systemd-resolved.service will result in a host name lookup for which
923 systemd-resolved.service needs to be started already. There are
924 multiple ways to work around this problem: pre-allocate the
925 "systemd-resolve" user on such systems, so that nss-ldap won't be
926 triggered; or use a different NSS package that doesn't do networking
927 in-process but provides a local asynchronous name cache; or configure
928 the NSS package to avoid lookups for UIDs in the range `pkg-config
929 systemd --variable=dynamicuidmin` … `pkg-config systemd
930 --variable=dynamicuidmax`, so that it does not consider itself
931 authoritative for the same UID range systemd allocates dynamic users
934 * The systemd-resolve tool has been renamed to resolvectl (it also
935 remains available under the old name, for compatibility), and its
936 interface is now verb-based, similar in style to the other <xyz>ctl
937 tools, such as systemctl or loginctl.
939 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
940 compatibility. It may be symlinked under the 'resolvconf' name, in
941 which case it will take arguments and input compatible with the
942 Debian and FreeBSD resolvconf tool.
944 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
945 where the system initially suspends, and after a timeout resumes and
948 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
949 set the client will only send a DUID as client identifier.
951 * The nss-systemd glibc NSS module will now enumerate dynamic users and
952 groups in effect. Previously, it could resolve UIDs/GIDs to user
953 names/groups and vice versa, but did not support enumeration.
955 * journald's Compress= configuration setting now optionally accepts a
956 byte threshold value. All journal objects larger than this threshold
957 will be compressed, smaller ones will not. Previously this threshold
958 was not configurable and set to 512.
960 * A new system.conf setting NoNewPrivileges= is now available which may
961 be used to turn off acquisition of new privileges system-wide
962 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
963 for all its children). Note that turning this option on means setuid
964 binaries and file system capabilities lose their special powers.
965 While turning on this option is a big step towards a more secure
966 system, doing so is likely to break numerous pre-existing UNIX tools,
967 in particular su and sudo.
969 * A new service systemd-time-sync-wait.service has been added. If
970 enabled it will delay the time-sync.target unit at boot until time
971 synchronization has been received from the network. This
972 functionality is useful on systems lacking a local RTC or where it is
973 acceptable that the boot process shall be delayed by external network
976 * When hibernating, systemd will now inform the kernel of the image
977 write offset, on kernels new enough to support this. This means swap
978 files should work for hibernation now.
980 * When loading unit files, systemd will now look for drop-in unit files
981 extensions in additional places. Previously, for a unit file name
982 "foo-bar-baz.service" it would look for dropin files in
983 "foo-bar-baz.service.d/*.conf". Now, it will also look in
984 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
985 service name truncated after all inner dashes. This scheme allows
986 writing drop-ins easily that apply to a whole set of unit files at
987 once. It's particularly useful for mount and slice units (as their
988 naming is prefix based), but is also useful for service and other
989 units, for packages that install multiple unit files at once,
990 following a strict naming regime of beginning the unit file name with
991 the package's name. Two new specifiers are now supported in unit
992 files to match this: %j and %J are replaced by the part of the unit
993 name following the last dash.
995 * Unit files and other configuration files that support specifier
996 expansion now understand another three new specifiers: %T and %V will
997 resolve to /tmp and /var/tmp respectively, or whatever temporary
998 directory has been set for the calling user. %E will expand to either
999 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
1001 * The ExecStart= lines of unit files are no longer required to
1002 reference absolute paths. If non-absolute paths are specified the
1003 specified binary name is searched within the service manager's
1004 built-in $PATH, which may be queried with 'systemd-path
1005 search-binaries-default'. It's generally recommended to continue to
1006 use absolute paths for all binaries specified in unit files.
1008 * Units gained a new load state "bad-setting", which is used when a
1009 unit file was loaded, but contained fatal errors which prevent it
1010 from being started (for example, a service unit has been defined
1011 lacking both ExecStart= and ExecStop= lines).
1013 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
1014 support alternative debuggers, for example lldb. The old name
1015 continues to be available however, for compatibility reasons. Use the
1016 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
1017 to pick an alternative debugger instead of the default gdb.
1019 * systemctl and the other tools will now output escape sequences that
1020 generate proper clickable hyperlinks in various terminal emulators
1021 where useful (for example, in the "systemctl status" output you can
1022 now click on the unit file name to quickly open it in the
1023 editor/viewer of your choice). Note that not all terminal emulators
1024 support this functionality yet, but many do. Unfortunately, the
1025 "less" pager doesn't support this yet, hence this functionality is
1026 currently automatically turned off when a pager is started (which
1027 happens quite often due to auto-paging). We hope to remove this
1028 limitation as soon as "less" learns these escape sequences. This new
1029 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
1030 environment variable. For details on these escape sequences see:
1031 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
1033 * networkd's .network files now support a new IPv6MTUBytes= option for
1034 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
1035 option in the [Route] section to configure the MTU to use for
1036 specific routes. It also gained support for configuration of the DHCP
1037 "UserClass" option through the new UserClass= setting. It gained
1038 three new options in the new [CAN] section for configuring CAN
1039 networks. The MULTICAST and ALLMULTI interface flags may now be
1040 controlled explicitly with the new Multicast= and AllMulticast=
1043 * networkd will now automatically make use of the kernel's route
1044 expiration feature, if it is available.
1046 * udevd's .link files now support setting the number of receive and
1047 transmit channels, using the RxChannels=, TxChannels=,
1048 OtherChannels=, CombinedChannels= settings.
1050 * Support for UDPSegmentationOffload= has been removed, given its
1051 limited support in hardware, and waning software support.
1053 * networkd's .netdev files now support creating "netdevsim" interfaces.
1055 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
1056 to query the unit belonging to a specific kernel control group.
1058 * systemd-analyze gained a new verb "cat-config", which may be used to
1059 dump the contents of any configuration file, with all its matching
1060 drop-in files added in, and honouring the usual search and masking
1061 logic applied to systemd configuration files. For example use
1062 "systemd-analyze cat-config systemd/system.conf" to get the complete
1063 system configuration file of systemd how it would be loaded by PID 1
1064 itself. Similar to this, various tools such as systemd-tmpfiles or
1065 systemd-sysusers, gained a new option "--cat-config", which does the
1066 corresponding operation for their own configuration settings. For
1067 example, "systemd-tmpfiles --cat-config" will now output the full
1068 list of tmpfiles.d/ lines in place.
1070 * timedatectl gained three new verbs: "show" shows bus properties of
1071 systemd-timedated, "timesync-status" shows the current NTP
1072 synchronization state of systemd-timesyncd, and "show-timesync"
1073 shows bus properties of systemd-timesyncd.
1075 * systemd-timesyncd gained a bus interface on which it exposes details
1078 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
1079 understood by systemd-timedated. It takes a colon-separated list of
1080 unit names of NTP client services. The list is used by
1081 "timedatectl set-ntp".
1083 * systemd-nspawn gained a new --rlimit= switch for setting initial
1084 resource limits for the container payload. There's a new switch
1085 --hostname= to explicitly override the container's hostname. A new
1086 --no-new-privileges= switch may be used to control the
1087 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
1088 --oom-score-adjust= switch controls the OOM scoring adjustment value
1089 for the payload. The new --cpu-affinity= switch controls the CPU
1090 affinity of the container payload. The new --resolv-conf= switch
1091 allows more detailed control of /etc/resolv.conf handling of the
1092 container. Similarly, the new --timezone= switch allows more detailed
1093 control of /etc/localtime handling of the container.
1095 * systemd-detect-virt gained a new --list switch, which will print a
1096 list of all currently known VM and container environments.
1098 * Support for "Portable Services" has been added, see
1099 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
1100 experimental, but this is expected to change soon. Reflecting this
1101 experimental state, the "portablectl" binary is not installed into
1102 /usr/bin yet. The binary has to be called with the full path
1103 /usr/lib/systemd/portablectl instead.
1105 * journalctl's and systemctl's -o switch now knows a new log output
1106 mode "with-unit". The output it generates is very similar to the
1107 regular "short" mode, but displays the unit name instead of the
1108 syslog tag for each log line. Also, the date is shown with timezone
1109 information. This mode is probably more useful than the classic
1110 "short" output mode for most purposes, except where pixel-perfect
1111 compatibility with classic /var/log/messages formatting is required.
1113 * A new --dump-bus-properties switch has been added to the systemd
1114 binary, which may be used to dump all supported D-Bus properties.
1115 (Options which are still supported, but are deprecated, are *not*
1118 * sd-bus gained a set of new calls:
1119 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
1120 enable/disable the "floating" state of a bus slot object,
1121 i.e. whether the slot object pins the bus it is allocated for into
1122 memory or if the bus slot object gets disconnected when the bus goes
1123 away. sd_bus_open_with_description(),
1124 sd_bus_open_user_with_description(),
1125 sd_bus_open_system_with_description() may be used to allocate bus
1126 objects and set their description string already during allocation.
1128 * sd-event gained support for watching inotify events from the event
1129 loop, in an efficient way, sharing inotify handles between multiple
1130 users. For this a new function sd_event_add_inotify() has been added.
1132 * sd-event and sd-bus gained support for calling special user-supplied
1133 destructor functions for userdata pointers associated with
1134 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
1135 functions sd_bus_slot_set_destroy_callback,
1136 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
1137 sd_bus_track_get_destroy_callback,
1138 sd_event_source_set_destroy_callback,
1139 sd_event_source_get_destroy_callback have been added.
1141 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
1143 * PID 1 will now automatically reschedule .timer units whenever the
1144 local timezone changes. (They previously got rescheduled
1145 automatically when the system clock changed.)
1147 * New documentation has been added to document cgroups delegation,
1148 portable services and the various code quality tools we have set up:
1150 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
1151 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
1152 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
1154 * The Boot Loader Specification has been added to the source tree.
1156 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
1158 While moving it into our source tree we have updated it and further
1159 changes are now accepted through the usual github PR workflow.
1161 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
1162 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
1163 earlier PAM modules. The data in these fields is used to initialize
1164 the session scope's resource properties. Thus external PAM modules
1165 may now configure per-session limits, for example sourced from
1166 external user databases.
1168 * socket units with Accept=yes will now maintain a "refused" counter in
1169 addition to the existing "accepted" counter, counting connections
1170 refused due to the enforced limits.
1172 * The "systemd-path search-binaries-default" command may now be use to
1173 query the default, built-in $PATH PID 1 will pass to the services it
1176 * A new unit file setting PrivateMounts= has been added. It's a boolean
1177 option. If enabled the unit's processes are invoked in their own file
1178 system namespace. Note that this behaviour is also implied if any
1179 other file system namespacing options (such as PrivateTmp=,
1180 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
1181 primarily useful for services that do not use any of the other file
1182 system namespacing options. One such service is systemd-udevd.service
1183 wher this is now used by default.
1185 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
1186 when the system is booted in UEFI "secure mode".
1188 * A new unit "system-update-pre.target" is added, which defines an
1189 optional synchronization point for offline system updates, as
1190 implemented by the pre-existing "system-update.target" unit. It
1191 allows ordering services before the service that executes the actual
1192 update process in a generic way.
1194 * Systemd now emits warnings whenever .include syntax is used.
1196 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
1197 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
1198 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
1199 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
1200 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
1201 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
1202 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
1203 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
1204 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
1205 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
1206 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
1207 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
1208 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
1209 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
1210 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
1211 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
1212 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
1213 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
1214 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
1215 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
1216 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
1217 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
1218 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
1219 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
1220 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
1221 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
1222 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
1223 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
1224 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1226 — Berlin, 2018-06-22
1230 * The MemoryAccounting= unit property now defaults to on. After
1231 discussions with the upstream control group maintainers we learnt
1232 that the negative impact of cgroup memory accounting on current
1233 kernels is finally relatively minimal, so that it should be safe to
1234 enable this by default without affecting system performance. Besides
1235 memory accounting only task accounting is turned on by default, all
1236 other forms of resource accounting (CPU, IO, IP) remain off for now,
1237 because it's not clear yet that their impact is small enough to move
1238 from opt-in to opt-out. We recommend downstreams to leave memory
1239 accounting on by default if kernel 4.14 or higher is primarily
1240 used. On very resource constrained systems or when support for old
1241 kernels is a necessity, -Dmemory-accounting-default=false can be used
1242 to revert this change.
1244 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
1245 %udev_rules_update) and the journal catalog (%journal_catalog_update)
1246 from the upgrade scriptlets of individual packages now do nothing.
1247 Transfiletriggers have been added which will perform those updates
1248 once at the end of the transaction.
1250 Similar transfiletriggers have been added to execute any sysctl.d
1251 and binfmt.d rules. Thus, it should be unnecessary to provide any
1252 scriptlets to execute this configuration from package installation
1255 * systemd-sysusers gained a mode where the configuration to execute is
1256 specified on the command line, but this configuration is not executed
1257 directly, but instead it is merged with the configuration on disk,
1258 and the result is executed. This is useful for package installation
1259 scripts which want to create the user before installing any files on
1260 disk (in case some of those files are owned by that user), while
1261 still allowing local admin overrides.
1263 This functionality is exposed to rpm scriptlets through a new
1264 %sysusers_create_package macro. Old %sysusers_create and
1265 %sysusers_create_inline macros are deprecated.
1267 A transfiletrigger for sysusers.d configuration is now installed,
1268 which means that it should be unnecessary to call systemd-sysusers from
1269 package installation scripts, unless the package installs any files
1270 owned by those newly-created users, in which case
1271 %sysusers_create_package should be used.
1273 * Analogous change has been done for systemd-tmpfiles: it gained a mode
1274 where the command-line configuration is merged with the configuration
1275 on disk. This is exposed as the new %tmpfiles_create_package macro,
1276 and %tmpfiles_create is deprecated. A transfiletrigger is installed
1277 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
1278 from package installation scripts.
1280 * sysusers.d configuration for a user may now also specify the group
1281 number, in addition to the user number ("u username 123:456"), or
1282 without the user number ("u username -:456").
1284 * Configution items for systemd-sysusers can now be specified as
1285 positional arguments when the new --inline switch is used.
1287 * The login shell of users created through sysusers.d may now be
1288 specified (previously, it was always /bin/sh for root and
1289 /sbin/nologin for other users).
1291 * systemd-analyze gained a new --global switch to look at global user
1292 configuration. It also gained a unit-paths verb to list the unit load
1293 paths that are compiled into systemd (which can be used with
1294 --systemd, --user, or --global).
1296 * udevadm trigger gained a new --settle/-w option to wait for any
1297 triggered events to finish (but just those, and not any other events
1298 which are triggered meanwhile).
1300 * The action that systemd-logind takes when the lid is closed and the
1301 machine is connected to external power can now be configured using
1302 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
1303 was determined by HandleLidSwitch=, and, for backwards compatibility,
1304 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
1306 * journalctl will periodically call sd_journal_process() to make it
1307 resilient against inotify queue overruns when journal files are
1308 rotated very quickly.
1310 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
1311 sd_bus_get_n_queued_write — may be used to check the number of
1312 pending bus messages.
1314 * systemd gained a new
1315 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
1316 which can be used to migrate foreign processes to scope and service
1317 units. The primary user for this new API is systemd itself: the
1318 systemd --user instance uses this call of the systemd --system
1319 instance to migrate processes if it itself gets the request to
1320 migrate processes and the kernel refuses this due to access
1321 restrictions. Thanks to this "systemd-run --scope --user …" works
1322 again in pure cgroup v2 environments when invoked from the user
1325 * A new TemporaryFileSystem= setting can be used to mask out part of
1326 the real file system tree with tmpfs mounts. This may be combined
1327 with BindPaths= and BindReadOnlyPaths= to hide files or directories
1328 not relevant to the unit, while still allowing some paths lower in
1329 the tree to be accessed.
1331 ProtectHome=tmpfs may now be used to hide user home and runtime
1332 directories from units, in a way that is mostly equivalent to
1333 "TemporaryFileSystem=/home /run/user /root".
1335 * Non-service units are now started with KeyringMode=shared by default.
1336 This means that mount and swapon and other mount tools have access
1337 to keys in the main keyring.
1339 * /sys/fs/bpf is now mounted automatically.
1341 * QNX virtualization is now detected by systemd-detect-virt and may
1342 be used in ConditionVirtualization=.
1344 * IPAccounting= may now be enabled also for slice units.
1346 * A new -Dsplit-bin= build configuration switch may be used to specify
1347 whether bin and sbin directories are merged, or if they should be
1348 included separately in $PATH and various listings of executable
1349 directories. The build configuration scripts will try to autodetect
1350 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
1351 system, but distributions are encouraged to configure this
1354 * A new -Dok-color= build configuration switch may be used to change
1355 the colour of "OK" status messages.
1357 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
1358 PrivateNetwork=yes was buggy in previous versions of systemd. This
1359 means that after the upgrade and daemon-reexec, any such units must
1362 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
1363 will not exclude read-only files owned by root from cleanup.
1365 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
1366 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
1367 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
1368 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
1369 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
1370 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
1371 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
1372 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
1373 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
1374 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
1375 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
1376 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
1377 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1378 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
1379 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
1380 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
1382 — Warsaw, 2018-03-05
1386 * Some keyboards come with a zoom see-saw or rocker which until now got
1387 mapped to the Linux "zoomin/out" keys in hwdb. However, these
1388 keycodes are not recognized by any major desktop. They now produce
1389 Up/Down key events so that they can be used for scrolling.
1391 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
1392 slightly: previously, if an argument was specified for lines of this
1393 type (i.e. the right-most column was set) this string was appended to
1394 existing files each time systemd-tmpfiles was run. This behaviour was
1395 different from what the documentation said, and not particularly
1396 useful, as repeated systemd-tmpfiles invocations would not be
1397 idempotent and grow such files without bounds. With this release
1398 behaviour has been altered to match what the documentation says:
1399 lines of this type only have an effect if the indicated files don't
1400 exist yet, and only then the argument string is written to the file.
1402 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
1403 systemd-tmpfiles behaviour: previously, read-only files owned by root
1404 were always excluded from the file "aging" algorithm (i.e. the
1405 automatic clean-up of directories like /tmp based on
1406 atime/mtime/ctime). We intend to drop this restriction, and age files
1407 by default even when owned by root and read-only. This behaviour was
1408 inherited from older tools, but there have been requests to remove
1409 it, and it's not obvious why this restriction was made in the first
1410 place. Please speak up now, if you are aware of software that reqires
1411 this behaviour, otherwise we'll remove the restriction in v238.
1413 * A new environment variable $SYSTEMD_OFFLINE is now understood by
1414 systemctl. It takes a boolean argument. If on, systemctl assumes it
1415 operates on an "offline" OS tree, and will not attempt to talk to the
1416 service manager. Previously, this mode was implicitly enabled if a
1417 chroot() environment was detected, and this new environment variable
1418 now provides explicit control.
1420 * .path and .socket units may now be created transiently, too.
1421 Previously only service, mount, automount and timer units were
1422 supported as transient units. The systemd-run tool has been updated
1423 to expose this new functionality, you may hence use it now to bind
1424 arbitrary commands to path or socket activation on-the-fly from the
1425 command line. Moreover, almost all properties are now exposed for the
1426 unit types that already supported transient operation.
1428 * The systemd-mount command gained support for a new --owner= parameter
1429 which takes a user name, which is then resolved and included in uid=
1430 and gid= mount options string of the file system to mount.
1432 * A new unit condition ConditionControlGroupController= has been added
1433 that checks whether a specific cgroup controller is available.
1435 * Unit files, udev's .link files, and systemd-networkd's .netdev and
1436 .network files all gained support for a new condition
1437 ConditionKernelVersion= for checking against specific kernel
1440 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
1441 support for configuring device flags in the Flags= setting. In the
1442 same files, the [Tunnel] section gained support for configuring
1443 AllowLocalRemote=. The [Route] section in .network files gained
1444 support for configuring InitialCongestionWindow=,
1445 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
1446 understands RapidCommit=.
1448 * systemd-networkd's DHCPv6 support gained support for Prefix
1451 * sd-bus gained support for a new "watch-bind" feature. When this
1452 feature is enabled, an sd_bus connection may be set up to connect to
1453 an AF_UNIX socket in the file system as soon as it is created. This
1454 functionality is useful for writing early-boot services that
1455 automatically connect to the system bus as soon as it is started,
1456 without ugly time-based polling. systemd-networkd and
1457 systemd-resolved have been updated to make use of this
1458 functionality. busctl exposes this functionality in a new
1459 --watch-bind= command line switch.
1461 * sd-bus will now optionally synthesize a local "Connected" signal as
1462 soon as a D-Bus connection is set up fully. This message mirrors the
1463 already existing "Disconnected" signal which is synthesized when the
1464 connection is terminated. This signal is generally useful but
1465 particularly handy in combination with the "watch-bind" feature
1466 described above. Synthesizing of this message has to be requested
1467 explicitly through the new API call sd_bus_set_connected_signal(). In
1468 addition a new call sd_bus_is_ready() has been added that checks
1469 whether a connection is fully set up (i.e. between the "Connected" and
1470 "Disconnected" signals).
1472 * sd-bus gained two new calls sd_bus_request_name_async() and
1473 sd_bus_release_name_async() for asynchronously registering bus
1474 names. Similar, there is now sd_bus_add_match_async() for installing
1475 a signal match asynchronously. All of systemd's own services have
1476 been updated to make use of these calls. Doing these operations
1477 asynchronously has two benefits: it reduces the risk of deadlocks in
1478 case of cyclic dependencies between bus services, and it speeds up
1479 service initialization since synchronization points for bus
1480 round-trips are removed.
1482 * sd-bus gained two new calls sd_bus_match_signal() and
1483 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
1484 and sd_bus_add_match_async() but instead of taking a D-Bus match
1485 string take match fields as normal function parameters.
1487 * sd-bus gained two new calls sd_bus_set_sender() and
1488 sd_bus_message_set_sender() for setting the sender name of outgoing
1489 messages (either for all outgoing messages or for just one specific
1490 one). These calls are only useful in direct connections as on
1491 brokered connections the broker fills in the sender anyway,
1492 overwriting whatever the client filled in.
1494 * sd-event gained a new pseudo-handle that may be specified on all API
1495 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
1496 used this refers to the default event loop object of the calling
1497 thread. Note however that this does not implicitly allocate one —
1498 which has to be done prior by using sd_event_default(). Similarly
1499 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
1500 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
1501 to the default bus of the specified type of the calling thread. Here
1502 too this does not implicitly allocate bus connection objects, this
1503 has to be done prior with sd_bus_default() and friends.
1505 * sd-event gained a new call pair
1506 sd_event_source_{get|set}_io_fd_own(). This may be used to request
1507 automatic closure of the file descriptor an IO event source watches
1508 when the event source is destroyed.
1510 * systemd-networkd gained support for natively configuring WireGuard
1513 * In previous versions systemd synthesized user records both for the
1514 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
1515 internally. In order to simplify distribution-wide renames of the
1516 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
1517 new transitional flag file has been added: if
1518 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
1519 user and group record within the systemd codebase is disabled.
1521 * systemd-notify gained a new --uid= option for selecting the source
1522 user/UID to use for notification messages sent to the service
1525 * journalctl gained a new --grep= option to list only entries in which
1526 the message matches a certain pattern. By default matching is case
1527 insensitive if the pattern is lowercase, and case sensitive
1528 otherwise. Option --case-sensitive=yes|no can be used to override
1529 this an specify case sensitivity or case insensitivity.
1531 * There's now a "systemd-analyze service-watchdogs" command for printing
1532 the current state of the service runtime watchdog, and optionally
1533 enabling or disabling the per-service watchdogs system-wide if given a
1534 boolean argument (i.e. the concept you configure in WatchdogSec=), for
1535 debugging purposes. There's also a kernel command line option
1536 systemd.service_watchdogs= for controlling the same.
1538 * Two new "log-level" and "log-target" options for systemd-analyze were
1539 added that merge the now deprecated get-log-level, set-log-level and
1540 get-log-target, set-log-target pairs. The deprecated options are still
1541 understood for backwards compatibility. The two new options print the
1542 current value when no arguments are given, and set them when a
1543 level/target is given as an argument.
1545 * sysusers.d's "u" lines now optionally accept both a UID and a GID
1546 specification, separated by a ":" character, in order to create users
1547 where UID and GID do not match.
1549 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
1550 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
1551 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
1552 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
1553 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
1554 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
1555 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
1556 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
1557 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
1558 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
1559 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
1560 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
1561 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
1562 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
1563 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
1564 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
1565 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
1566 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
1567 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
1568 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
1575 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
1576 in v235 has been extended to also set the dummy.ko module option
1577 numdummies=0, preventing the kernel from automatically creating
1578 dummy0. All dummy interfaces must now be explicitly created.
1580 * Unknown '%' specifiers in configuration files are now rejected. This
1581 applies to units and tmpfiles.d configuration. Any percent characters
1582 that are followed by a letter or digit that are not supposed to be
1583 interpreted as the beginning of a specifier should be escaped by
1584 doubling ("%%"). (So "size=5%" is still accepted, as well as
1585 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
1586 valid specifiers today.)
1588 * systemd-resolved now maintains a new dynamic
1589 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
1590 recommended to make /etc/resolv.conf a symlink to it. This file
1591 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
1592 includes dynamically acquired search domains, achieving more correct
1593 DNS resolution by software that bypasses local DNS APIs such as NSS.
1595 * The "uaccess" udev tag has been dropped from /dev/kvm and
1596 /dev/dri/renderD*. These devices now have the 0666 permissions by
1597 default (but this may be changed at build-time). /dev/dri/renderD*
1598 will now be owned by the "render" group along with /dev/kfd.
1600 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
1601 systemd-journal-gatewayd.service and
1602 systemd-journal-upload.service. This means "nss-systemd" must be
1603 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
1604 services are resolved properly.
1606 * In /etc/fstab two new mount options are now understood:
1607 x-systemd.makefs and x-systemd.growfs. The former has the effect that
1608 the configured file system is formatted before it is mounted, the
1609 latter that the file system is resized to the full block device size
1610 after it is mounted (i.e. if the file system is smaller than the
1611 partition it resides on, it's grown). This is similar to the fsck
1612 logic in /etc/fstab, and pulls in systemd-makefs@.service and
1613 systemd-growfs@.service as necessary, similar to
1614 systemd-fsck@.service. Resizing is currently only supported on ext4
1617 * In systemd-networkd, the IPv6 RA logic now optionally may announce
1618 DNS server and domain information.
1620 * Support for the LUKS2 on-disk format for encrypted partitions has
1621 been added. This requires libcryptsetup2 during compilation and
1624 * The systemd --user instance will now signal "readiness" when its
1625 basic.target unit has been reached, instead of when the run queue ran
1626 empty for the first time.
1628 * Tmpfiles.d with user configuration are now also supported.
1629 systemd-tmpfiles gained a new --user switch, and snippets placed in
1630 ~/.config/user-tmpfiles.d/ and corresponding directories will be
1631 executed by systemd-tmpfiles --user running in the new
1632 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
1633 running in the user session.
1635 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
1636 %S resolves to the top-level state directory (/var/lib for the system
1637 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
1638 top-level cache directory (/var/cache for the system instance,
1639 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
1640 logs directory (/var/log for the system instance,
1641 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
1642 existing %t specifier, that resolves to the top-level runtime
1643 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
1646 * journalctl learnt a new parameter --output-fields= for limiting the
1647 set of journal fields to output in verbose and JSON output modes.
1649 * systemd-timesyncd's configuration file gained a new option
1650 RootDistanceMaxSec= for setting the maximum root distance of servers
1651 it'll use, as well as the new options PollIntervalMinSec= and
1652 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
1654 * bootctl gained a new command "list" for listing all available boot
1655 menu items on systems that follow the boot loader specification.
1657 * systemctl gained a new --dry-run switch that shows what would be done
1658 instead of doing it, and is currently supported by the shutdown and
1661 * ConditionSecurity= can now detect the TOMOYO security module.
1663 * Unit file [Install] sections are now also respected in unit drop-in
1664 files. This is intended to be used by drop-ins under /usr/lib/.
1666 * systemd-firstboot may now also set the initial keyboard mapping.
1668 * Udev "changed" events for devices which are exposed as systemd
1669 .device units are now propagated to units specified in
1670 ReloadPropagatedFrom= as reload requests.
1672 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
1673 unit template name (i.e. a name in the form of 'foobar@.service',
1674 without the instance component between the '@' and - the '.'), then
1675 the escaped sysfs path of the device is automatically used as the
1678 * SystemCallFilter= in unit files has been extended so that an "errno"
1679 can be specified individually for each system call. Example:
1680 SystemCallFilter=~uname:EILSEQ.
1682 * The cgroup delegation logic has been substantially updated. Delegate=
1683 now optionally takes a list of controllers (instead of a boolean, as
1684 before), which lists the controllers to delegate at least.
1686 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
1688 * A new LogLevelMax= setting configures the maximum log level any
1689 process of the service may log at (i.e. anything with a lesser
1690 priority than what is specified is automatically dropped). A new
1691 LogExtraFields= setting allows configuration of additional journal
1692 fields to attach to all log records generated by any of the unit's
1695 * New StandardInputData= and StandardInputText= settings along with the
1696 new option StandardInput=data may be used to configure textual or
1697 binary data that shall be passed to the executed service process via
1698 standard input, encoded in-line in the unit file.
1700 * StandardInput=, StandardOutput= and StandardError= may now be used to
1701 connect stdin/stdout/stderr of executed processes directly with a
1702 file or AF_UNIX socket in the file system, using the new "file:" option.
1704 * A new unit file option CollectMode= has been added, that allows
1705 tweaking the garbage collection logic for units. It may be used to
1706 tell systemd to garbage collect units that have failed automatically
1707 (normally it only GCs units that exited successfully). systemd-run
1708 and systemd-mount expose this new functionality with a new -G option.
1710 * "machinectl bind" may now be used to bind mount non-directories
1711 (i.e. regularfiles, devices, fifos, sockets).
1713 * systemd-analyze gained a new verb "calendar" for validating and
1714 testing calendar time specifications to use for OnCalendar= in timer
1715 units. Besides validating the expression it will calculate the next
1716 time the specified expression would elapse.
1718 * In addition to the pre-existing FailureAction= unit file setting
1719 there's now SuccessAction=, for configuring a shutdown action to
1720 execute when a unit completes successfully. This is useful in
1721 particular inside containers that shall terminate after some workload
1722 has been completed. Also, both options are now supported for all unit
1723 types, not just services.
1725 * networkds's IP rule support gained two new options
1726 IncomingInterface= and OutgoingInterface= for configuring the incoming
1727 and outgoing interfaces of configured rules. systemd-networkd also
1728 gained support for "vxcan" network devices.
1730 * networkd gained a new setting RequiredForOnline=, taking a
1731 boolean. If set, systemd-wait-online will take it into consideration
1732 when determining that the system is up, otherwise it will ignore the
1733 interface for this purpose.
1735 * The sd_notify() protocol gained support for a new operation: with
1736 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
1737 store again, ahead of POLLHUP or POLLERR when they are removed
1740 * A new document doc/UIDS-GIDS.md has been added to the source tree,
1741 that documents the UID/GID range and assignment assumptions and
1742 requirements of systemd.
1744 * The watchdog device PID 1 will ping may now be configured through the
1745 WatchdogDevice= configuration file setting, or by setting the
1746 systemd.watchdog_service= kernel commandline option.
1748 * systemd-resolved's gained support for registering DNS-SD services on
1749 the local network using MulticastDNS. Services may either be
1750 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
1751 the same dir below /run, /usr/lib), or through its D-Bus API.
1753 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
1754 extend the effective start, runtime, and stop time. The service must
1755 continue to send EXTEND_TIMEOUT_USEC within the period specified to
1756 prevent the service manager from making the service as timedout.
1758 * systemd-resolved's DNSSEC support gained support for RFC 8080
1759 (Ed25519 keys and signatures).
1761 * The systemd-resolve command line tool gained a new set of options
1762 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
1763 --set-nta= and --revert to configure per-interface DNS configuration
1764 dynamically during runtime. It's useful for pushing DNS information
1765 into systemd-resolved from DNS hook scripts that various interface
1766 managing software supports (such as pppd).
1768 * systemd-nspawn gained a new --network-namespace-path= command line
1769 option, which may be used to make a container join an existing
1770 network namespace, by specifying a path to a "netns" file.
1772 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
1773 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
1774 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
1775 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
1776 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
1777 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
1778 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
1779 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
1780 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
1781 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
1782 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
1783 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
1784 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
1785 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
1786 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
1787 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
1788 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
1789 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1790 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
1791 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
1792 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
1793 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
1794 Jędrzejewski-Szmek, Zeal Jagannatha
1796 — Berlin, 2017-12-14
1800 * INCOMPATIBILITY: systemd-logind.service and other long-running
1801 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
1802 communication with the outside. This generally improves security of
1803 the system, and is in almost all cases a safe and good choice, as
1804 these services do not and should not provide any network-facing
1805 functionality. However, systemd-logind uses the glibc NSS API to
1806 query the user database. This creates problems on systems where NSS
1807 is set up to directly consult network services for user database
1808 lookups. In particular, this creates incompatibilities with the
1809 "nss-nis" module, which attempts to directly contact the NIS/YP
1810 network servers it is configured for, and will now consistently
1811 fail. In such cases, it is possible to turn off IP sandboxing for
1812 systemd-logind.service (set IPAddressDeny= in its [Service] section
1813 to the empty string, via a .d/ unit file drop-in). Downstream
1814 distributions might want to update their nss-nis packaging to include
1815 such a drop-in snippet, accordingly, to hide this incompatibility
1816 from the user. Another option is to make use of glibc's nscd service
1817 to proxy such network requests through a privilege-separated, minimal
1818 local caching daemon, or to switch to more modern technologies such
1819 sssd, whose NSS hook-ups generally do not involve direct network
1820 access. In general, we think it's definitely time to question the
1821 implementation choices of nss-nis, i.e. whether it's a good idea
1822 today to embed a network-facing loadable module into all local
1823 processes that need to query the user database, including the most
1824 trivial and benign ones, such as "ls". For more details about
1825 IPAddressDeny= see below.
1827 * A new modprobe.d drop-in is now shipped by default that sets the
1828 bonding module option max_bonds=0. This overrides the kernel default,
1829 to avoid conflicts and ambiguity as to whether or not bond0 should be
1830 managed by systemd-networkd or not. This resolves multiple issues
1831 with bond0 properties not being applied, when bond0 is configured
1832 with systemd-networkd. Distributors may choose to not package this,
1833 however in that case users will be prevented from correctly managing
1834 bond0 interface using systemd-networkd.
1836 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
1837 which print the logging level and target of the system manager. They
1838 complement the existing "set-log-level" and "set-log-target" verbs
1839 used to change those values.
1841 * journald.conf gained a new boolean setting ReadKMsg= which defaults
1842 to on. If turned off kernel log messages will not be read by
1843 systemd-journald or included in the logs. It also gained a new
1844 setting LineMax= for configuring the maximum line length in
1845 STDOUT/STDERR log streams. The new default for this value is 48K, up
1846 from the previous hardcoded 2048.
1848 * A new unit setting RuntimeDirectoryPreserve= has been added, which
1849 allows more detailed control of what to do with a runtime directory
1850 configured with RuntimeDirectory= (i.e. a directory below /run or
1851 $XDG_RUNTIME_DIR) after a unit is stopped.
1853 * The RuntimeDirectory= setting for units gained support for creating
1854 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
1855 one top-level directory.
1857 * Units gained new options StateDirectory=, CacheDirectory=,
1858 LogsDirectory= and ConfigurationDirectory= which are closely related
1859 to RuntimeDirectory= but manage per-service directories below
1860 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
1861 possible to write unit files which when activated automatically gain
1862 properly owned service specific directories in these locations, thus
1863 making unit files self-contained and increasing compatibility with
1864 stateless systems and factory reset where /etc or /var are
1865 unpopulated at boot. Matching these new settings there's also
1866 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
1867 ConfigurationDirectoryMode= for configuring the access mode of these
1868 directories. These settings are particularly useful in combination
1869 with DynamicUser=yes as they provide secure, properly-owned,
1870 writable, and stateful locations for storage, excluded from the
1871 sandbox that such services live in otherwise.
1873 * Automake support has been removed from this release. systemd is now
1876 * systemd-journald will now aggressively cache client metadata during
1877 runtime, speeding up log write performance under pressure. This comes
1878 at a small price though: as much of the metadata is read
1879 asynchronously from /proc/ (and isn't implicitly attached to log
1880 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
1881 metadata stored alongside a log entry might be slightly
1882 out-of-date. Previously it could only be slightly newer than the log
1883 message. The time window is small however, and given that the kernel
1884 is unlikely to be improved anytime soon in this regard, this appears
1887 * nss-myhostname/systemd-resolved will now by default synthesize an
1888 A/AAAA resource record for the "_gateway" hostname, pointing to the
1889 current default IP gateway. Previously it did that for the "gateway"
1890 name, hampering adoption, as some distributions wanted to leave that
1891 host name open for local use. The old behaviour may still be
1892 requested at build time.
1894 * systemd-networkd's [Address] section in .network files gained a new
1895 Scope= setting for configuring the IP address scope. The [Network]
1896 section gained a new boolean setting ConfigureWithoutCarrier= that
1897 tells systemd-networkd to ignore link sensing when configuring the
1898 device. The [DHCP] section gained a new Anonymize= boolean option for
1899 turning on a number of options suggested in RFC 7844. A new
1900 [RoutingPolicyRule] section has been added for configuring the IP
1901 routing policy. The [Route] section has gained support for a new
1902 Type= setting which permits configuring
1903 blackhole/unreachable/prohibit routes.
1905 * The [VRF] section in .netdev files gained a new Table= setting for
1906 configuring the routing table to use. The [Tunnel] section gained a
1907 new Independent= boolean field for configuring tunnels independent of
1908 an underlying network interface. The [Bridge] section gained a new
1909 GroupForwardMask= option for configuration of propagation of link
1910 local frames between bridge ports.
1912 * The WakeOnLan= setting in .link files gained support for a number of
1913 new modes. A new TCP6SegmentationOffload= setting has been added for
1914 configuring TCP/IPv6 hardware segmentation offload.
1916 * The IPv6 RA sender implementation may now optionally send out RDNSS
1917 and RDNSSL records to supply DNS configuration to peers.
1919 * systemd-nspawn gained support for a new --system-call-filter= command
1920 line option for adding and removing entries in the default system
1921 call filter it applies. Moreover systemd-nspawn has been changed to
1922 implement a system call whitelist instead of a blacklist.
1924 * systemd-run gained support for a new --pipe command line option. If
1925 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1926 are directly passed on to the activated transient service
1927 executable. This allows invoking arbitrary processes as systemd
1928 services (for example to take benefit of dependency management,
1929 accounting management, resource management or log management that is
1930 done automatically for services) — while still allowing them to be
1931 integrated in a classic UNIX shell pipeline.
1933 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1934 using ReloadPropagationTo= is configured, a reload is now propagated
1935 to configured units. (Previously this was only done on explicitly
1936 requested reloads, using "systemctl reload" or an equivalent
1939 * For each service unit a restart counter is now kept: it is increased
1940 each time the service is restarted due to Restart=, and may be
1941 queried using "systemctl show -p NRestarts …".
1943 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1944 @signal and @timer have been added, for usage with SystemCallFilter=
1945 in unit files and the new --system-call-filter= command line option
1946 of systemd-nspawn (see above).
1948 * ExecStart= lines in unit files gained two new modifiers: when a
1949 command line is prefixed with "!" the command will be executed as
1950 configured, except for the credentials applied by
1951 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1952 "+", but does still apply namespacing options unlike "+". There's
1953 also "!!" now, which is mostly identical, but becomes a NOP on
1954 systems that support ambient capabilities. This is useful to write
1955 unit files that work with ambient capabilities where possible but
1956 automatically fall back to traditional privilege dropping mechanisms
1957 on systems where this is not supported.
1959 * ListenNetlink= settings in socket units now support RDMA netlink
1962 * A new unit file setting LockPersonality= has been added which permits
1963 locking down the chosen execution domain ("personality") of a service
1966 * A new special target "getty-pre.target" has been added, which is
1967 ordered before all text logins, and may be used to order services
1968 before textual logins acquire access to the console.
1970 * systemd will now attempt to load the virtio-rng.ko kernel module very
1971 early on if a VM environment supporting this is detected. This should
1972 improve entropy during early boot in virtualized environments.
1974 * A _netdev option is now supported in /etc/crypttab that operates in a
1975 similar way as the same option in /etc/fstab: it permits configuring
1976 encrypted devices that need to be ordered after the network is up.
1977 Following this logic, two new special targets
1978 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
1979 added that are to cryptsetup.target what remote-fs.target and
1980 remote-fs-pre.target are to local-fs.target.
1982 * Service units gained a new UnsetEnvironment= setting which permits
1983 unsetting specific environment variables for services that are
1984 normally passed to it (for example in order to mask out locale
1985 settings for specific services that can't deal with it).
1987 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1988 traffic accounting (packet count as well as byte count) is done for
1989 the service, and shown as part of "systemctl status" or "systemd-run
1992 * Service units acquired two new options IPAddressAllow= and
1993 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1994 for configuring a simple IP access control list for all sockets of
1995 the unit. These options are available also on .slice and .socket
1996 units, permitting flexible access list configuration for individual
1997 services as well as groups of services (as defined by a slice unit),
1998 including system-wide. Note that IP ACLs configured this way are
1999 enforced on every single IPv4 and IPv6 socket created by any process
2000 of the service unit, and apply to ingress as well as egress traffic.
2002 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
2003 structured log message is generated each time the unit is stopped,
2004 containing information about the consumed resources of this
2007 * A new setting KeyringMode= has been added to unit files, which may be
2008 used to control how the kernel keyring is set up for executed
2011 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
2012 "systemctl kexec" and "systemctl exit" are now always asynchronous in
2013 behaviour (that is: these commands return immediately after the
2014 operation was enqueued instead of waiting for the operation to
2015 complete). Previously, "systemctl poweroff" and "systemctl reboot"
2016 were asynchronous on systems using systemd-logind (i.e. almost
2017 always, and like they were on sysvinit), and the other three commands
2018 were unconditionally synchronous. With this release this is cleaned
2019 up, and callers will see the same asynchronous behaviour on all
2020 systems for all five operations.
2022 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
2025 * .timer units now accept calendar specifications in other timezones
2026 than UTC or the local timezone.
2028 * The tmpfiles snippet var.conf has been changed to create
2029 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
2030 the "utmp" group already, and it appears to be generally understood
2031 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
2032 databases. Previously this was implemented correctly for all these
2033 databases excepts btmp, which has been opened up like this now
2034 too. Note that while the other databases are world-readable
2035 (i.e. 0644), btmp is not and remains more restrictive.
2037 * The systemd-resolve tool gained a new --reset-server-features
2038 switch. When invoked like this systemd-resolved will forget
2039 everything it learnt about the features supported by the configured
2040 upstream DNS servers, and restarts the feature probing logic on the
2041 next resolver look-up for them at the highest feature level
2044 * The status dump systemd-resolved sends to the logs upon receiving
2045 SIGUSR1 now also includes information about all DNS servers it is
2046 configured to use, and the features levels it probed for them.
2048 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
2049 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
2050 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
2051 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
2052 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
2053 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
2054 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
2055 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
2056 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
2057 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
2058 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
2059 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
2060 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
2061 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
2062 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
2063 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
2064 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
2065 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
2066 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
2067 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2069 — Berlin, 2017-10-06
2073 * Meson is now supported as build system in addition to Automake. It is
2074 our plan to remove Automake in one of our next releases, so that
2075 Meson becomes our exclusive build system. Hence, please start using
2076 the Meson build system in your downstream packaging. There's plenty
2077 of documentation around how to use Meson, the extremely brief
2080 ./autogen.sh && ./configure && make && sudo make install
2084 meson build && ninja -C build && sudo ninja -C build install
2086 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
2087 which permits configuring a timeout on the time a job is
2088 running. This is particularly useful for setting timeouts on jobs for
2091 * Unit files gained two new options ConditionUser= and ConditionGroup=
2092 for conditionalizing units based on the identity of the user/group
2093 running a systemd user instance.
2095 * systemd-networkd now understands a new FlowLabel= setting in the
2096 [VXLAN] section of .network files, as well as a Priority= in
2097 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
2098 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
2099 gained support for configuration of GENEVE links, and IPv6 address
2100 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
2102 * .link files now understand a new Port= setting.
2104 * systemd-networkd's DHCP support gained support for DHCP option 119
2105 (domain search list).
2107 * systemd-networkd gained support for serving IPv6 address ranges using
2108 the Router Advertisement protocol. The new .network configuration
2109 section [IPv6Prefix] may be used to configure the ranges to
2110 serve. This is implemented based on a new, minimal, native server
2111 implementation of RA.
2113 * journalctl's --output= switch gained support for a new parameter
2114 "short-iso-precise" for a mode where timestamps are shown as precise
2117 * systemd-udevd's "net_id" builtin may now generate stable network
2118 interface names from IBM PowerVM VIO devices as well as ACPI platform
2121 * MulticastDNS support in systemd-resolved may now be explicitly
2122 enabled/disabled using the new MulticastDNS= configuration file
2125 * systemd-resolved may now optionally use libidn2 instead of the libidn
2126 for processing internationalized domain names. Support for libidn2
2127 should be considered experimental and should not be enabled by
2130 * "machinectl pull-tar" and related call may now do verification of
2131 downloaded images using SUSE-style .sha256 checksum files in addition
2132 to the already existing support for validating using Ubuntu-style
2135 * sd-bus gained support for a new sd_bus_message_appendv() call which
2136 is va_list equivalent of sd_bus_message_append().
2138 * sd-boot gained support for validating images using SHIM/MOK.
2140 * The SMACK code learnt support for "onlycap".
2142 * systemd-mount --umount is now much smarter in figuring out how to
2143 properly unmount a device given its mount or device path.
2145 * The code to call libnss_dns as a fallback from libnss_resolve when
2146 the communication with systemd-resolved fails was removed. This
2147 fallback was redundant and interfered with the [!UNAVAIL=return]
2148 suffix. See nss-resolve(8) for the recommended configuration.
2150 * systemd-logind may now be restarted without losing state. It stores
2151 the file descriptors for devices it manages in the system manager
2152 using the FDSTORE= mechanism. Please note that further changes in
2153 other components may be required to make use of this (for example
2154 Xorg has code to listen for stops of systemd-logind and terminate
2155 itself when logind is stopped or restarted, in order to avoid using
2156 stale file descriptors for graphical devices, which is now
2157 counterproductive and must be reverted in order for restarts of
2158 systemd-logind to be safe. See
2159 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
2161 * All kernel-install plugins are called with the environment variable
2162 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
2163 /etc/machine-id. If the machine ID could not be determined,
2164 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
2165 anything in the entry directory (passed as the second argument) if
2166 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatiblity, a
2167 temporary directory is passed as the entry directory and removed
2168 after all the plugins exit.
2170 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
2171 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
2172 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
2173 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
2174 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
2175 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
2176 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
2177 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
2178 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
2179 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
2180 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
2181 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
2182 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
2183 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
2184 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
2185 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
2186 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
2187 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
2188 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
2189 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
2190 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
2191 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
2192 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
2193 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
2194 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
2195 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
2196 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
2197 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
2198 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
2201 — Berlin, 2017-07-12
2205 * The "hybrid" control group mode has been modified to improve
2206 compatibility with "legacy" cgroups-v1 setups. Specifically, the
2207 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
2208 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
2209 cgroups-v1 hierarchy), the only externally visible change being that
2210 the cgroups-v2 hierarchy is also mounted, to
2211 /sys/fs/cgroup/unified. This should provide a large degree of
2212 compatibility with "legacy" cgroups-v1, while taking benefit of the
2213 better management capabilities of cgroups-v2.
2215 * The default control group setup mode may be selected both a boot-time
2216 via a set of kernel command line parameters (specifically:
2217 systemd.unified_cgroup_hierarchy= and
2218 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
2219 default selected on the configure command line
2220 (--with-default-hierarchy=). The upstream default is "hybrid"
2221 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
2222 this will change in a future systemd version to be "unified" (pure
2223 cgroups-v2 mode). The third option for the compile time option is
2224 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
2225 distributions to default to "hybrid" mode for release distributions,
2226 starting with v233. We recommend "unified" for development
2227 distributions (specifically: distributions such as Fedora's rawhide)
2228 as that's where things are headed in the long run. Use "legacy" for
2229 greatest stability and compatibility only.
2231 * Note one current limitation of "unified" and "hybrid" control group
2232 setup modes: the kernel currently does not permit the systemd --user
2233 instance (i.e. unprivileged code) to migrate processes between two
2234 disconnected cgroup subtrees, even if both are managed and owned by
2235 the user. This effectively means "systemd-run --user --scope" doesn't
2236 work when invoked from outside of any "systemd --user" service or
2237 scope. Specifically, it is not supported from session scopes. We are
2238 working on fixing this in a future systemd version. (See #3388 for
2239 further details about this.)
2241 * DBus policy files are now installed into /usr rather than /etc. Make
2242 sure your system has dbus >= 1.9.18 running before upgrading to this
2243 version, or override the install path with --with-dbuspolicydir= .
2245 * All python scripts shipped with systemd (specifically: the various
2246 tests written in Python) now require Python 3.
2248 * systemd unit tests can now run standalone (without the source or
2249 build directories), and can be installed into /usr/lib/systemd/tests/
2250 with 'make install-tests'.
2252 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
2253 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
2256 * Support for the %c, %r, %R specifiers in unit files has been
2257 removed. Specifiers are not supposed to be dependent on configuration
2258 in the unit file itself (so that they resolve the same regardless
2259 where used in the unit files), but these specifiers were influenced
2260 by the Slice= option.
2262 * The shell invoked by debug-shell.service now defaults to /bin/sh in
2263 all cases. If distributions want to use a different shell for this
2264 purpose (for example Fedora's /sbin/sushell) they need to specify
2265 this explicitly at configure time using --with-debug-shell=.
2267 * The confirmation spawn prompt has been reworked to offer the
2270 (c)ontinue, proceed without asking anymore
2271 (D)ump, show the state of the unit
2272 (f)ail, don't execute the command and pretend it failed
2274 (i)nfo, show a short summary of the unit
2275 (j)obs, show jobs that are in progress
2276 (s)kip, don't execute the command and pretend it succeeded
2277 (y)es, execute the command
2279 The 'n' choice for the confirmation spawn prompt has been removed,
2280 because its meaning was confusing.
2282 The prompt may now also be redirected to an alternative console by
2283 specifying the console as parameter to systemd.confirm_spawn=.
2285 * Services of Type=notify require a READY=1 notification to be sent
2286 during startup. If no such message is sent, the service now fails,
2287 even if the main process exited with a successful exit code.
2289 * Services that fail to start up correctly now always have their
2290 ExecStopPost= commands executed. Previously, they'd enter "failed"
2291 state directly, without executing these commands.
2293 * The option MulticastDNS= of network configuration files has acquired
2294 an actual implementation. With MulticastDNS=yes a host can resolve
2295 names of remote hosts and reply to mDNS A and AAAA requests.
2297 * When units are about to be started an additional check is now done to
2298 ensure that all dependencies of type BindsTo= (when used in
2299 combination with After=) have been started.
2301 * systemd-analyze gained a new verb "syscall-filter" which shows which
2302 system call groups are defined for the SystemCallFilter= unit file
2303 setting, and which system calls they contain.
2305 * A new system call filter group "@filesystem" has been added,
2306 consisting of various file system related system calls. Group
2307 "@reboot" has been added, covering reboot, kexec and shutdown related
2308 calls. Finally, group "@swap" has been added covering swap
2309 configuration related calls.
2311 * A new unit file option RestrictNamespaces= has been added that may be
2312 used to restrict access to the various process namespace types the
2313 Linux kernel provides. Specifically, it may be used to take away the
2314 right for a service unit to create additional file system, network,
2315 user, and other namespaces. This sandboxing option is particularly
2316 relevant due to the high amount of recently discovered namespacing
2317 related vulnerabilities in the kernel.
2319 * systemd-udev's .link files gained support for a new AutoNegotiation=
2320 setting for configuring Ethernet auto-negotiation.
2322 * systemd-networkd's .network files gained support for a new
2323 ListenPort= setting in the [DHCP] section to explicitly configure the
2324 UDP client port the DHCP client shall listen on.
2326 * .network files gained a new Unmanaged= boolean setting for explicitly
2327 excluding one or more interfaces from management by systemd-networkd.
2329 * The systemd-networkd ProxyARP= option has been renamed to
2330 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
2331 renamed to ReduceARPProxy=. The old names continue to be available
2334 * systemd-networkd gained support for configuring IPv6 Proxy NDP
2335 addresses via the new IPv6ProxyNDPAddress= .network file setting.
2337 * systemd-networkd's bonding device support gained support for two new
2338 configuration options ActiveSlave= and PrimarySlave=.
2340 * The various options in the [Match] section of .network files gained
2341 support for negative matching.
2343 * New systemd-specific mount options are now understood in /etc/fstab:
2345 x-systemd.mount-timeout= may be used to configure the maximum
2346 permitted runtime of the mount command.
2348 x-systemd.device-bound may be set to bind a mount point to its
2349 backing device unit, in order to automatically remove a mount point
2350 if its backing device is unplugged. This option may also be
2351 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
2352 on the block device, which is now automatically set for all CDROM
2353 drives, so that mounted CDs are automatically unmounted when they are
2354 removed from the drive.
2356 x-systemd.after= and x-systemd.before= may be used to explicitly
2357 order a mount after or before another unit or mount point.
2359 * Enqueued start jobs for device units are now automatically garbage
2360 collected if there are no jobs waiting for them anymore.
2362 * systemctl list-jobs gained two new switches: with --after, for every
2363 queued job the jobs it's waiting for are shown; with --before the
2364 jobs which it's blocking are shown.
2366 * systemd-nspawn gained support for ephemeral boots from disk images
2367 (or in other words: --ephemeral and --image= may now be
2368 combined). Moreover, ephemeral boots are now supported for normal
2369 directories, even if the backing file system is not btrfs. Of course,
2370 if the file system does not support file system snapshots or
2371 reflinks, the initial copy operation will be relatively expensive, but
2372 this should still be suitable for many use cases.
2374 * Calendar time specifications in .timer units now support
2375 specifications relative to the end of a month by using "~" instead of
2376 "-" as separator between month and day. For example, "*-02~03" means
2377 "the third last day in February". In addition a new syntax for
2378 repeated events has been added using the "/" character. For example,
2379 "9..17/2:00" means "every two hours from 9am to 5pm".
2381 * systemd-socket-proxyd gained a new parameter --connections-max= for
2382 configuring the maximum number of concurrent connections.
2384 * sd-id128 gained a new API for generating unique IDs for the host in a
2385 way that does not leak the machine ID. Specifically,
2386 sd_id128_get_machine_app_specific() derives an ID based on the
2387 machine ID a in well-defined, non-reversible, stable way. This is
2388 useful whenever an identifier for the host is needed but where the
2389 identifier shall not be useful to identify the system beyond the
2390 scope of the application itself. (Internally this uses HMAC-SHA256 as
2391 keyed hash function using the machine ID as input.)
2393 * NotifyAccess= gained a new supported value "exec". When set
2394 notifications are accepted from all processes systemd itself invoked,
2395 including all control processes.
2397 * .nspawn files gained support for defining overlay mounts using the
2398 Overlay= and OverlayReadOnly= options. Previously this functionality
2399 was only available on the systemd-nspawn command line.
2401 * systemd-nspawn's --bind= and --overlay= options gained support for
2402 bind/overlay mounts whose source lies within the container tree by
2403 prefixing the source path with "+".
2405 * systemd-nspawn's --bind= and --overlay= options gained support for
2406 automatically allocating a temporary source directory in /var/tmp
2407 that is removed when the container dies. Specifically, if the source
2408 directory is specified as empty string this mechanism is selected. An
2409 example usage is --overlay=+/var::/var, which creates an overlay
2410 mount based on the original /var contained in the image, overlayed
2411 with a temporary directory in the host's /var/tmp. This way changes
2412 to /var are automatically flushed when the container shuts down.
2414 * systemd-nspawn --image= option does now permit raw file system block
2415 devices (in addition to images containing partition tables, as
2418 * The disk image dissection logic in systemd-nspawn gained support for
2419 automatically setting up LUKS encrypted as well as Verity protected
2420 partitions. When a container is booted from an encrypted image the
2421 passphrase is queried at start-up time. When a container with Verity
2422 data is started, the root hash is search in a ".roothash" file
2423 accompanying the disk image (alternatively, pass the root hash via
2424 the new --root-hash= command line option).
2426 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
2427 be used to dissect disk images the same way as systemd-nspawn does
2428 it, following the Bootable Partition Specification. It may even be
2429 used to mount disk images with complex partition setups (including
2430 LUKS and Verity partitions) to a local host directory, in order to
2431 inspect them. This tool is not considered public API (yet), and is
2432 thus not installed into /usr/bin. Please do not rely on its
2433 existence, since it might go away or be changed in later systemd
2436 * A new generator "systemd-verity-generator" has been added, similar in
2437 style to "systemd-cryptsetup-generator", permitting automatic setup of
2438 Verity root partitions when systemd boots up. In order to make use of
2439 this your partition setup should follow the Discoverable Partitions
2440 Specification, and the GPT partition ID of the root file system
2441 partition should be identical to the upper 128bit of the Verity root
2442 hash. The GPT partition ID of the Verity partition protecting it
2443 should be the lower 128bit of the Verity root hash. If the partition
2444 image follows this model it is sufficient to specify a single
2445 "roothash=" kernel command line argument to both configure which root
2446 image and verity partition to use as well as the root hash for
2447 it. Note that systemd-nspawn's Verity support follows the same
2448 semantics, meaning that disk images with proper Verity data in place
2449 may be booted in containers with systemd-nspawn as well as on
2450 physical systems via the verity generator. Also note that the "mkosi"
2451 tool available at https://github.com/systemd/mkosi has been updated
2452 to generate Verity protected disk images following this scheme. In
2453 fact, it has been updated to generate disk images that optionally
2454 implement a complete UEFI SecureBoot trust chain, involving a signed
2455 kernel and initrd image that incorporates such a root hash as well as
2456 a Verity-enabled root partition.
2458 * The hardware database (hwdb) udev supports has been updated to carry
2459 accelerometer quirks.
2461 * All system services are now run with a fresh kernel keyring set up
2462 for them. The invocation ID is stored by default in it, thus
2463 providing a safe, non-overridable way to determine the invocation
2466 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
2467 options for bind mounting arbitrary paths in a service-specific
2468 way. When these options are used, arbitrary host or service files and
2469 directories may be mounted to arbitrary locations in the service's
2472 * Documentation has been added that lists all of systemd's low-level
2473 environment variables:
2475 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
2477 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
2478 whether a specific socket file descriptor matches a specified socket
2481 * systemd-firstboot has been updated to check for the
2482 systemd.firstboot= kernel command line option. It accepts a boolean
2483 and when set to false the first boot questions are skipped.
2485 * systemd-fstab-generator has been updated to check for the
2486 systemd.volatile= kernel command line option, which either takes an
2487 optional boolean parameter or the special value "state". If used the
2488 system may be booted in a "volatile" boot mode. Specifically,
2489 "systemd.volatile" is used, the root directory will be mounted as
2490 tmpfs, and only /usr is mounted from the actual root file system. If
2491 "systemd.volatile=state" is used, the root directory will be mounted
2492 as usual, but /var is mounted as tmpfs. This concept provides similar
2493 functionality as systemd-nspawn's --volatile= option, but provides it
2494 on physical boots. Use this option for implementing stateless
2495 systems, or testing systems with all state and/or configuration reset
2496 to the defaults. (Note though that many distributions are not
2497 prepared to boot up without a populated /etc or /var, though.)
2499 * systemd-gpt-auto-generator gained support for LUKS encrypted root
2500 partitions. Previously it only supported LUKS encrypted partitions
2501 for all other uses, except for the root partition itself.
2503 * Socket units gained support for listening on AF_VSOCK sockets for
2504 communication in virtualized QEMU environments.
2506 * The "configure" script gained a new option --with-fallback-hostname=
2507 for specifying the fallback hostname to use if none is configured in
2508 /etc/hostname. For example, by specifying
2509 --with-fallback-hostname=fedora it is possible to default to a
2510 hostname of "fedora" on pristine installations.
2512 * systemd-cgls gained support for a new --unit= switch for listing only
2513 the control groups of a specific unit. Similar --user-unit= has been
2514 added for listing only the control groups of a specific user unit.
2516 * systemd-mount gained a new --umount switch for unmounting a mount or
2517 automount point (and all mount/automount points below it).
2519 * systemd will now refuse full configuration reloads (via systemctl
2520 daemon-reload and related calls) unless at least 16MiB of free space
2521 are available in /run. This is a safety precaution in order to ensure
2522 that generators can safely operate after the reload completed.
2524 * A new unit file option RootImage= has been added, which has a similar
2525 effect as RootDirectory= but mounts the service's root directory from
2526 a disk image instead of plain directory. This logic reuses the same
2527 image dissection and mount logic that systemd-nspawn already uses,
2528 and hence supports any disk images systemd-nspawn supports, including
2529 those following the Discoverable Partition Specification, as well as
2530 Verity enabled images. This option enables systemd to run system
2531 services directly off disk images acting as resource bundles,
2532 possibly even including full integrity data.
2534 * A new MountAPIVFS= unit file option has been added, taking a boolean
2535 argument. If enabled /proc, /sys and /dev (collectively called the
2536 "API VFS") will be mounted for the service. This is only relevant if
2537 RootDirectory= or RootImage= is used for the service, as these mounts
2538 are of course in place in the host mount namespace anyway.
2540 * systemd-nspawn gained support for a new --pivot-root= switch. If
2541 specified the root directory within the container image is pivoted to
2542 the specified mount point, while the original root disk is moved to a
2543 different place. This option enables booting of ostree images
2544 directly with systemd-nspawn.
2546 * The systemd build scripts will no longer complain if the NTP server
2547 addresses are not changed from the defaults. Google now supports
2548 these NTP servers officially. We still recommend downstreams to
2549 properly register an NTP pool with the NTP pool project though.
2551 * coredumpctl gained a new "--reverse" option for printing the list
2552 of coredumps in reverse order.
2554 * coredumpctl will now show additional information about truncated and
2555 inaccessible coredumps, as well as coredumps that are still being
2556 processed. It also gained a new --quiet switch for suppressing
2557 additional informational message in its output.
2559 * coredumpctl gained support for only showing coredumps newer and/or
2560 older than specific timestamps, using the new --since= and --until=
2561 options, reminiscent of journalctl's options by the same name.
2563 * The systemd-coredump logic has been improved so that it may be reused
2564 to collect backtraces in non-compiled languages, for example in
2565 scripting languages such as Python.
2567 * machinectl will now show the UID shift of local containers, if user
2568 namespacing is enabled for them.
2570 * systemd will now optionally run "environment generator" binaries at
2571 configuration load time. They may be used to add environment
2572 variables to the environment block passed to services invoked. One
2573 user environment generator is shipped by default that sets up
2574 environment variables based on files dropped into /etc/environment.d
2575 and ~/.config/environment.d/.
2577 * systemd-resolved now includes the new, recently published 2017 DNSSEC
2580 * hostnamed has been updated to report a new chassis type of
2581 "convertible" to cover "foldable" laptops that can both act as a
2582 tablet and as a laptop, such as various Lenovo Yoga devices.
2584 Contributions from: Adrián López, Alexander Galanin, Alexander
2585 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
2586 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
2587 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
2588 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
2589 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
2590 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
2591 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
2592 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
2593 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
2594 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
2595 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
2596 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
2597 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
2598 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
2599 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
2600 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
2601 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
2602 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
2603 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
2604 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
2605 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
2606 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
2607 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
2608 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
2609 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
2610 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
2611 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
2614 — Berlin, 2017-03-01
2618 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
2619 RestrictAddressFamilies= enabled. These sandboxing options should
2620 generally be compatible with the various external udev call-out
2621 binaries we are aware of, however there may be exceptions, in
2622 particular when exotic languages for these call-outs are used. In
2623 this case, consider turning off these settings locally.
2625 * The new RemoveIPC= option can be used to remove IPC objects owned by
2626 the user or group of a service when that service exits.
2628 * The new ProtectKernelModules= option can be used to disable explicit
2629 load and unload operations of kernel modules by a service. In
2630 addition access to /usr/lib/modules is removed if this option is set.
2632 * ProtectSystem= option gained a new value "strict", which causes the
2633 whole file system tree with the exception of /dev, /proc, and /sys,
2634 to be remounted read-only for a service.
2636 * The new ProtectKernelTunables= option can be used to disable
2637 modification of configuration files in /sys and /proc by a service.
2638 Various directories and files are remounted read-only, so access is
2639 restricted even if the file permissions would allow it.
2641 * The new ProtectControlGroups= option can be used to disable write
2642 access by a service to /sys/fs/cgroup.
2644 * Various systemd services have been hardened with
2645 ProtectKernelTunables=yes, ProtectControlGroups=yes,
2646 RestrictAddressFamilies=.
2648 * Support for dynamically creating users for the lifetime of a service
2649 has been added. If DynamicUser=yes is specified, user and group IDs
2650 will be allocated from the range 61184..65519 for the lifetime of the
2651 service. They can be resolved using the new nss-systemd.so NSS
2652 module. The module must be enabled in /etc/nsswitch.conf. Services
2653 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
2654 any resources allocated by the service will be cleaned up when the
2655 service exits. They also have ProtectHome=read-only and
2656 ProtectSystem=strict enabled, so they are not able to make any
2657 permanent modifications to the system.
2659 * The nss-systemd module also always resolves root and nobody, making
2660 it possible to have no /etc/passwd or /etc/group files in minimal
2661 container or chroot environments.
2663 * Services may be started with their own user namespace using the new
2664 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
2665 under which the service is running are mapped. All other users are
2668 * Support for the cgroup namespace has been added to systemd-nspawn. If
2669 supported by kernel, the container system started by systemd-nspawn
2670 will have its own view of the cgroup hierarchy. This new behaviour
2671 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
2673 * The new MemorySwapMax= option can be used to limit the maximum swap
2674 usage under the unified cgroup hierarchy.
2676 * Support for the CPU controller in the unified cgroup hierarchy has
2677 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
2678 options. This controller requires out-of-tree patches for the kernel
2679 and the support is provisional.
2681 * Mount and automount units may now be created transiently
2682 (i.e. dynamically at runtime via the bus API, instead of requiring
2683 unit files in the file system).
2685 * systemd-mount is a new tool which may mount file systems – much like
2686 mount(8), optionally pulling in additional dependencies through
2687 transient .mount and .automount units. For example, this tool
2688 automatically runs fsck on a backing block device before mounting,
2689 and allows the automount logic to be used dynamically from the
2690 command line for establishing mount points. This tool is particularly
2691 useful when dealing with removable media, as it will ensure fsck is
2692 run – if necessary – before the first access and that the file system
2693 is quickly unmounted after each access by utilizing the automount
2694 logic. This maximizes the chance that the file system on the
2695 removable media stays in a clean state, and if it isn't in a clean
2696 state is fixed automatically.
2698 * LazyUnmount=yes option for mount units has been added to expose the
2699 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
2702 * /efi will be used as the mount point of the EFI boot partition, if
2703 the directory is present, and the mount point was not configured
2704 through other means (e.g. fstab). If /efi directory does not exist,
2705 /boot will be used as before. This makes it easier to automatically
2706 mount the EFI partition on systems where /boot is used for something
2709 * When operating on GPT disk images for containers, systemd-nspawn will
2710 now mount the ESP to /boot or /efi according to the same rules as PID
2711 1 running on a host. This allows tools like "bootctl" to operate
2712 correctly within such containers, in order to make container images
2713 bootable on physical systems.
2715 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
2717 * Two new user session targets have been added to support running
2718 graphical sessions under the systemd --user instance:
2719 graphical-session.target and graphical-session-pre.target. See
2720 systemd.special(7) for a description of how those targets should be
2723 * The vconsole initialization code has been significantly reworked to
2724 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
2725 support unicode keymaps. Font and keymap configuration will now be
2726 copied to all allocated virtual consoles.
2728 * FreeBSD's bhyve virtualization is now detected.
2730 * Information recorded in the journal for core dumps now includes the
2731 contents of /proc/mountinfo and the command line of the process at
2732 the top of the process hierarchy (which is usually the init process
2735 * systemd-journal-gatewayd learned the --directory= option to serve
2736 files from the specified location.
2738 * journalctl --root=… can be used to peruse the journal in the
2739 /var/log/ directories inside of a container tree. This is similar to
2740 the existing --machine= option, but does not require the container to
2743 * The hardware database has been extended to support
2744 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
2747 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
2748 specify the click rate for mice which include a horizontal wheel with
2749 a click rate that is different than the one for the vertical wheel.
2751 * systemd-run gained a new --wait option that makes service execution
2752 synchronous. (Specifically, the command will not return until the
2753 specified service binary exited.)
2755 * systemctl gained a new --wait option that causes the start command to
2756 wait until the units being started have terminated again.
2758 * A new journal output mode "short-full" has been added which displays
2759 timestamps with abbreviated English day names and adds a timezone
2760 suffix. Those timestamps include more information than the default
2761 "short" output mode, and can be passed directly to journalctl's
2762 --since= and --until= options.
2764 * /etc/resolv.conf will be bind-mounted into containers started by
2765 systemd-nspawn, if possible, so any changes to resolv.conf contents
2766 are automatically propagated to the container.
2768 * The number of instances for socket-activated services originating
2769 from a single IP address can be limited with
2770 MaxConnectionsPerSource=, extending the existing setting of
2773 * systemd-networkd gained support for vcan ("Virtual CAN") interface
2776 * .netdev and .network configuration can now be extended through
2779 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
2780 Segmentation Offload, Generic Receive Offload, Large Receive Offload
2781 can be enabled and disabled using the new UDPSegmentationOffload=,
2782 TCPSegmentationOffload=, GenericSegmentationOffload=,
2783 GenericReceiveOffload=, LargeReceiveOffload= options in the
2784 [Link] section of .link files.
2786 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
2787 Port VLAN ID can be configured for bridge devices using the new STP=,
2788 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
2789 section of .netdev files.
2791 * The route table to which routes received over DHCP or RA should be
2792 added can be configured with the new RouteTable= option in the [DHCP]
2793 and [IPv6AcceptRA] sections of .network files.
2795 * The Address Resolution Protocol can be disabled on links managed by
2796 systemd-networkd using the ARP=no setting in the [Link] section of
2799 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
2800 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
2801 encode information about the result and exit codes of the current
2802 service runtime cycle.
2804 * systemd-sysctl will now configure kernel parameters in the order
2805 they occur in the configuration files. This matches what sysctl
2806 has been traditionally doing.
2808 * kernel-install "plugins" that are executed to perform various
2809 tasks after a new kernel is added and before an old one is removed
2810 can now return a special value to terminate the procedure and
2811 prevent any later plugins from running.
2813 * Journald's SplitMode=login setting has been deprecated. It has been
2814 removed from documentation, and its use is discouraged. In a future
2815 release it will be completely removed, and made equivalent to current
2816 default of SplitMode=uid.
2818 * Storage=both option setting in /etc/systemd/coredump.conf has been
2819 removed. With fast LZ4 compression storing the core dump twice is not
2822 * The --share-system systemd-nspawn option has been replaced with an
2823 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
2824 this functionality is discouraged. In addition the variables
2825 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
2826 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
2827 individual namespaces.
2829 * "machinectl list" now shows the IP address of running containers in
2830 the output, as well as OS release information.
2832 * "loginctl list" now shows the TTY of each session in the output.
2834 * sd-bus gained new API calls sd_bus_track_set_recursive(),
2835 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
2836 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
2837 tracking objects in a "recursive" mode, where a single client can be
2838 counted multiple times, if it takes multiple references.
2840 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
2841 sd_bus_get_exit_on_disconnect(). They may be used to make a
2842 process using sd-bus automatically exit if the bus connection is
2845 * Bus clients of the service manager may now "pin" loaded units into
2846 memory, by taking an explicit reference on them. This is useful to
2847 ensure the client can retrieve runtime data about the service even
2848 after the service completed execution. Taking such a reference is
2849 available only for privileged clients and should be helpful to watch
2850 running services in a race-free manner, and in particular collect
2851 information about exit statuses and results.
2853 * The nss-resolve module has been changed to strictly return UNAVAIL
2854 when communication via D-Bus with resolved failed, and NOTFOUND when
2855 a lookup completed but was negative. This means it is now possible to
2856 neatly configure fallbacks using nsswitch.conf result checking
2857 expressions. Taking benefit of this, the new recommended
2858 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
2860 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
2862 * A new setting CtrlAltDelBurstAction= has been added to
2863 /etc/systemd/system.conf which may be used to configure the precise
2864 behaviour if the user on the console presses Ctrl-Alt-Del more often
2865 than 7 times in 2s. Previously this would unconditionally result in
2866 an expedited, immediate reboot. With this new setting the precise
2867 operation may be configured in more detail, and also turned off
2870 * In .netdev files two new settings RemoteChecksumTx= and
2871 RemoteChecksumRx= are now understood that permit configuring the
2872 remote checksumming logic for VXLAN networks.
2874 * The service manager learnt a new "invocation ID" concept for invoked
2875 services. Each runtime cycle of a service will get a new invocation
2876 ID (a 128bit random UUID) assigned that identifies the current
2877 run of the service uniquely and globally. A new invocation ID
2878 is generated each time a service starts up. The journal will store
2879 the invocation ID of a service along with any logged messages, thus
2880 making the invocation ID useful for matching the online runtime of a
2881 service with the offline log data it generated in a safe way without
2882 relying on synchronized timestamps. In many ways this new service
2883 invocation ID concept is similar to the kernel's boot ID concept that
2884 uniquely and globally identifies the runtime of each boot. The
2885 invocation ID of a service is passed to the service itself via an
2886 environment variable ($INVOCATION_ID). A new bus call
2887 GetUnitByInvocationID() has been added that is similar to GetUnit()
2888 but instead of retrieving the bus path for a unit by its name
2889 retrieves it by its invocation ID. The returned path is valid only as
2890 long as the passed invocation ID is current.
2892 * systemd-resolved gained a new "DNSStubListener" setting in
2893 resolved.conf. It either takes a boolean value or the special values
2894 "udp" and "tcp", and configures whether to enable the stub DNS
2895 listener on 127.0.0.53:53.
2897 * IP addresses configured via networkd may now carry additional
2898 configuration settings supported by the kernel. New options include:
2899 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
2900 PrefixRoute=, AutoJoin=.
2902 * The PAM configuration fragment file for "user@.service" shipped with
2903 systemd (i.e. the --user instance of systemd) has been stripped to
2904 the minimum necessary to make the system boot. Previously, it
2905 contained Fedora-specific stanzas that did not apply to other
2906 distributions. It is expected that downstream distributions add
2907 additional configuration lines, matching their needs to this file,
2908 using it only as rough template of what systemd itself needs. Note
2909 that this reduced fragment does not even include an invocation of
2910 pam_limits which most distributions probably want to add, even though
2911 systemd itself does not need it. (There's also the new build time
2912 option --with-pamconfdir=no to disable installation of the PAM
2915 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2916 capability is now also dropped from its set (in addition to
2917 CAP_SYS_MKNOD as before).
2919 * In service unit files it is now possible to connect a specific named
2920 file descriptor with stdin/stdout/stdout of an executed service. The
2921 name may be specified in matching .socket units using the
2922 FileDescriptorName= setting.
2924 * A number of journal settings may now be configured on the kernel
2925 command line. Specifically, the following options are now understood:
2926 systemd.journald.max_level_console=,
2927 systemd.journald.max_level_store=,
2928 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2929 systemd.journald.max_level_wall=.
2931 * "systemctl is-enabled --full" will now show by which symlinks a unit
2932 file is enabled in the unit dependency tree.
2934 * Support for VeraCrypt encrypted partitions has been added to the
2935 "cryptsetup" logic and /etc/crypttab.
2937 * systemd-detect-virt gained support for a new --private-users switch
2938 that checks whether the invoking processes are running inside a user
2939 namespace. Similar, a new special value "private-users" for the
2940 existing ConditionVirtualization= setting has been added, permitting
2941 skipping of specific units in user namespace environments.
2943 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2944 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2945 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2946 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2947 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2948 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2949 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2950 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2951 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2952 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2953 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2954 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2955 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2956 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2957 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2958 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2959 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2960 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2961 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2962 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2963 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2964 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2965 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2966 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2967 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2968 Jędrzejewski-Szmek, Zeal Jagannatha
2970 — Santa Fe, 2016-11-03
2974 * In service units the various ExecXYZ= settings have been extended
2975 with an additional special character as first argument of the
2976 assigned value: if the character '+' is used the specified command
2977 line it will be run with full privileges, regardless of User=,
2978 Group=, CapabilityBoundingSet= and similar options. The effect is
2979 similar to the existing PermissionsStartOnly= option, but allows
2980 configuration of this concept for each executed command line
2983 * Services may now alter the service watchdog timeout at runtime by
2984 sending a WATCHDOG_USEC= message via sd_notify().
2986 * MemoryLimit= and related unit settings now optionally take percentage
2987 specifications. The percentage is taken relative to the amount of
2988 physical memory in the system (or in case of containers, the assigned
2989 amount of memory). This allows scaling service resources neatly with
2990 the amount of RAM available on the system. Similarly, systemd-logind's
2991 RuntimeDirectorySize= option now also optionally takes percentage
2994 * In similar fashion TasksMax= takes percentage values now, too. The
2995 value is taken relative to the configured maximum number of processes
2996 on the system. The per-service task maximum has been changed to 15%
2997 using this functionality. (Effectively this is an increase of 512 →
2998 4915 for service units, given the kernel's default pid_max setting.)
3000 * Calendar time specifications in .timer units now understand a ".."
3001 syntax for time ranges. Example: "4..7:10" may now be used for
3002 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
3005 * The InaccessableDirectories=, ReadOnlyDirectories= and
3006 ReadWriteDirectories= unit file settings have been renamed to
3007 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
3008 applied to all kinds of file nodes, and not just directories, with
3009 the exception of symlinks. Specifically these settings may now be
3010 used on block and character device nodes, UNIX sockets and FIFOS as
3011 well as regular files. The old names of these settings remain
3012 available for compatibility.
3014 * systemd will now log about all service processes it kills forcibly
3015 (using SIGKILL) because they remained after the clean shutdown phase
3016 of the service completed. This should help identifying services that
3017 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
3018 systemd-logind's configuration a similar log message is generated for
3019 processes killed at the end of each session due to this setting.
3021 * systemd will now set the $JOURNAL_STREAM environment variable for all
3022 services whose stdout/stderr are connected to the Journal (which
3023 effectively means by default: all services). The variable contains
3024 the device and inode number of the file descriptor used for
3025 stdout/stderr. This may be used by invoked programs to detect whether
3026 their stdout/stderr is connected to the Journal, in which case they
3027 can switch over to direct Journal communication, thus being able to
3028 pass extended, structured metadata along with their log messages. As
3029 one example, this is now used by glib's logging primitives.
3031 * When using systemd's default tmp.mount unit for /tmp, the mount point
3032 will now be established with the "nosuid" and "nodev" options. This
3033 avoids privilege escalation attacks that put traps and exploits into
3034 /tmp. However, this might cause problems if you e. g. put container
3035 images or overlays into /tmp; if you need this, override tmp.mount's
3036 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
3039 * systemd now supports the "memory" cgroup controller also on
3042 * The systemd-cgtop tool now optionally takes a control group path as
3043 command line argument. If specified, the control group list shown is
3044 limited to subgroups of that group.
3046 * The SystemCallFilter= unit file setting gained support for
3047 pre-defined, named system call filter sets. For example
3048 SystemCallFilter=@clock is now an effective way to make all clock
3049 changing-related system calls unavailable to a service. A number of
3050 similar pre-defined groups are defined. Writing system call filters
3051 for system services is simplified substantially with this new
3052 concept. Accordingly, all of systemd's own, long-running services now
3053 enable system call filtering based on this, by default.
3055 * A new service setting MemoryDenyWriteExecute= has been added, taking
3056 a boolean value. If turned on, a service may no longer create memory
3057 mappings that are writable and executable at the same time. This
3058 enhances security for services where this is enabled as it becomes
3059 harder to dynamically write and then execute memory in exploited
3060 service processes. This option has been enabled for all of systemd's
3061 own long-running services.
3063 * A new RestrictRealtime= service setting has been added, taking a
3064 boolean argument. If set the service's processes may no longer
3065 acquire realtime scheduling. This improves security as realtime
3066 scheduling may otherwise be used to easily freeze the system.
3068 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
3069 value. This may be used for requesting that the system manager inside
3070 of the container reports start-up completion to nspawn which then
3071 propagates this notification further to the service manager
3072 supervising nspawn itself. A related option NotifyReady= in .nspawn
3073 files has been added too. This functionality allows ordering of the
3074 start-up of multiple containers using the usual systemd ordering
3077 * machinectl gained a new command "stop" that is an alias for
3080 * systemd-resolved gained support for contacting DNS servers on
3081 link-local IPv6 addresses.
3083 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
3084 its caches. A method call for requesting the same operation has been
3085 added to the bus API too, and is made available via "systemd-resolve
3088 * systemd-resolve gained a new --status switch. If passed a brief
3089 summary of the used DNS configuration with per-interface information
3092 * resolved.conf gained a new Cache= boolean option, defaulting to
3093 on. If turned off local DNS caching is disabled. This comes with a
3094 performance penalty in particular when DNSSEC is enabled. Note that
3095 resolved disables its internal caching implicitly anyway, when the
3096 configured DNS server is on a host-local IP address such as ::1 or
3097 127.0.0.1, thus automatically avoiding double local caching.
3099 * systemd-resolved now listens on the local IP address 127.0.0.53:53
3100 for DNS requests. This improves compatibility with local programs
3101 that do not use the libc NSS or systemd-resolved's bus APIs for name
3102 resolution. This minimal DNS service is only available to local
3103 programs and does not implement the full DNS protocol, but enough to
3104 cover local DNS clients. A new, static resolv.conf file, listing just
3105 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
3106 now recommended to make /etc/resolv.conf a symlink to this file in
3107 order to route all DNS lookups to systemd-resolved, regardless if
3108 done via NSS, the bus API or raw DNS packets. Note that this local
3109 DNS service is not as fully featured as the libc NSS or
3110 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
3111 used to deliver link-local address information (as this implies
3112 sending a local interface index along), LLMNR/mDNS support via this
3113 interface is severely restricted. It is thus strongly recommended for
3114 all applications to use the libc NSS API or native systemd-resolved
3117 * systemd-networkd's bridge support learned a new setting
3118 VLANFiltering= for controlling VLAN filtering. Moreover a new section
3119 in .network files has been added for configuring VLAN bridging in
3120 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
3122 * systemd-networkd's IPv6 Router Advertisement code now makes use of
3123 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
3124 now be acquired without relying on DHCPv6. Two new options
3125 UseDomains= and UseDNS= have been added to configure this behaviour.
3127 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
3128 renamed IPv6AcceptRA=, without altering its behaviour. The old
3129 setting name remains available for compatibility reasons.
3131 * The systemd-networkd VTI/VTI6 tunneling support gained new options
3132 Key=, InputKey= and OutputKey=.
3134 * systemd-networkd gained support for VRF ("Virtual Routing Function")
3135 interface configuration.
3137 * "systemctl edit" may now be used to create new unit files by
3138 specifying the --force switch.
3140 * sd-event gained a new function sd_event_get_iteration() for
3141 requesting the current iteration counter of the event loop. It starts
3142 at zero and is increased by one with each event loop iteration.
3144 * A new rpm macro %systemd_ordering is provided by the macros.systemd
3145 file. It can be used in lieu of %systemd_requires in packages which
3146 don't use any systemd functionality and are intended to be installed
3147 in minimal containers without systemd present. This macro provides
3148 ordering dependencies to ensure that if the package is installed in
3149 the same rpm transaction as systemd, systemd will be installed before
3150 the scriptlets for the package are executed, allowing unit presets
3153 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
3154 been added to simplify packaging of generators.
3156 * The os-release file gained VERSION_CODENAME field for the
3157 distribution nickname (e.g. VERSION_CODENAME=woody).
3159 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
3160 can be set to disable parsing of metadata and the creation
3161 of persistent symlinks for that device.
3163 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
3164 to make them available to logged-in users has been reverted.
3166 * Much of the common code of the various systemd components is now
3167 built into an internal shared library libsystemd-shared-231.so
3168 (incorporating the systemd version number in the name, to be updated
3169 with future releases) that the components link to. This should
3170 decrease systemd footprint both in memory during runtime and on
3171 disk. Note that the shared library is not for public use, and is
3172 neither API nor ABI stable, but is likely to change with every new
3173 released update. Packagers need to make sure that binaries
3174 linking to libsystemd-shared.so are updated in step with the
3177 * Configuration for "mkosi" is now part of the systemd
3178 repository. mkosi is a tool to easily build legacy-free OS images,
3179 and is available on github: https://github.com/systemd/mkosi. If
3180 "mkosi" is invoked in the build tree a new raw OS image is generated
3181 incorporating the systemd sources currently being worked on and a
3182 clean, fresh distribution installation. The generated OS image may be
3183 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
3184 UEFI PC. This functionality is particularly useful to easily test
3185 local changes made to systemd in a pristine, defined environment. See
3186 doc/HACKING for details.
3188 * configure learned the --with-support-url= option to specify the
3189 distribution's bugtracker.
3191 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
3192 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
3193 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
3194 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
3195 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
3196 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
3197 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
3198 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
3199 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
3200 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
3201 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
3202 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
3203 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
3204 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
3205 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
3206 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
3207 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
3208 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
3209 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
3211 — Berlin, 2016-07-25
3215 * DNSSEC is now turned on by default in systemd-resolved (in
3216 "allow-downgrade" mode), but may be turned off during compile time by
3217 passing "--with-default-dnssec=no" to "configure" (and of course,
3218 during runtime with DNSSEC= in resolved.conf). We recommend
3219 downstreams to leave this on at least during development cycles and
3220 report any issues with the DNSSEC logic upstream. We are very
3221 interested in collecting feedback about the DNSSEC validator and its
3222 limitations in the wild. Note however, that DNSSEC support is
3223 probably nothing downstreams should turn on in stable distros just
3224 yet, as it might create incompatibilities with a few DNS servers and
3225 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
3226 automatically whenever we detect such incompatible setups, but there
3227 might be systems we do not cover yet. Hence: please help us testing
3228 the DNSSEC code, leave this on where you can, report back, but then
3229 again don't consider turning this on in your stable, LTS or
3230 production release just yet. (Note that you have to enable
3231 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
3232 and its DNSSEC mode for host name resolution from local
3235 * systemd-resolve conveniently resolves DANE records with the --tlsa
3236 option and OPENPGPKEY records with the --openpgp option. It also
3237 supports dumping raw DNS record data via the new --raw= switch.
3239 * systemd-logind will now by default terminate user processes that are
3240 part of the user session scope unit (session-XX.scope) when the user
3241 logs out. This behavior is controlled by the KillUserProcesses=
3242 setting in logind.conf, and the previous default of "no" is now
3243 changed to "yes". This means that user sessions will be properly
3244 cleaned up after, but additional steps are necessary to allow
3245 intentionally long-running processes to survive logout.
3247 While the user is logged in at least once, user@.service is running,
3248 and any service that should survive the end of any individual login
3249 session can be started at a user service or scope using systemd-run.
3250 systemd-run(1) man page has been extended with an example which shows
3251 how to run screen in a scope unit underneath user@.service. The same
3252 command works for tmux.
3254 After the user logs out of all sessions, user@.service will be
3255 terminated too, by default, unless the user has "lingering" enabled.
3256 To effectively allow users to run long-term tasks even if they are
3257 logged out, lingering must be enabled for them. See loginctl(1) for
3258 details. The default polkit policy was modified to allow users to
3259 set lingering for themselves without authentication.
3261 Previous defaults can be restored at compile time by the
3262 --without-kill-user-processes option to "configure".
3264 * systemd-logind gained new configuration settings SessionsMax= and
3265 InhibitorsMax=, both with a default of 8192. It will not register new
3266 user sessions or inhibitors above this limit.
3268 * systemd-logind will now reload configuration on SIGHUP.
3270 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
3271 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
3272 enable. Also, support for the "io" cgroup controller in the unified
3273 hierarchy has been added, so that the "memory", "pids" and "io" are
3274 now the controllers that are supported on the unified hierarchy.
3276 WARNING: it is not possible to use previous systemd versions with
3277 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
3278 is necessary to also update systemd in the initramfs if using the
3279 unified hierarchy. An updated SELinux policy is also required.
3281 * LLDP support has been extended, and both passive (receive-only) and
3282 active (sender) modes are supported. Passive mode ("routers-only") is
3283 enabled by default in systemd-networkd. Active LLDP mode is enabled
3284 by default for containers on the internal network. The "networkctl
3285 lldp" command may be used to list information gathered. "networkctl
3286 status" will also show basic LLDP information on connected peers now.
3288 * The IAID and DUID unique identifier sent in DHCP requests may now be
3289 configured for the system and each .network file managed by
3290 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
3292 * systemd-networkd gained support for configuring proxy ARP support for
3293 each interface, via the ProxyArp= setting in .network files. It also
3294 gained support for configuring the multicast querier feature of
3295 bridge devices, via the new MulticastQuerier= setting in .netdev
3296 files. Similarly, snooping on the IGMP traffic can be controlled
3297 via the new setting MulticastSnooping=.
3299 A new setting PreferredLifetime= has been added for addresses
3300 configured in .network file to configure the lifetime intended for an
3303 The systemd-networkd DHCP server gained the option EmitRouter=, which
3304 defaults to yes, to configure whether the DHCP Option 3 (Router)
3307 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
3308 systemd-socket-activate and installed into /usr/bin. It is now fully
3311 * systemd-journald now uses separate threads to flush changes to disk
3312 when closing journal files, thus reducing impact of slow disk I/O on
3313 logging performance.
3315 * The sd-journal API gained two new calls
3316 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
3317 can be used to open journal files using file descriptors instead of
3318 file or directory paths. sd_journal_open_container() has been
3319 deprecated, sd_journal_open_directory_fd() should be used instead
3320 with the flag SD_JOURNAL_OS_ROOT.
3322 * journalctl learned a new output mode "-o short-unix" that outputs log
3323 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
3324 UTC). It also gained support for a new --no-hostname setting to
3325 suppress the hostname column in the family of "short" output modes.
3327 * systemd-ask-password now optionally skips printing of the password to
3328 stdout with --no-output which can be useful in scripts.
3330 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
3331 (devices tagged with ID_MAKER_TOOL) are now tagged with
3332 "uaccess" and are available to logged in users.
3334 * The DeviceAllow= unit setting now supports specifiers (with "%").
3336 * "systemctl show" gained a new --value switch, which allows print a
3337 only the contents of a specific unit property, without also printing
3338 the property's name. Similar support was added to "show*" verbs
3339 of loginctl and machinectl that output "key=value" lists.
3341 * A new unit type "generated" was added for files dynamically generated
3342 by generator tools. Similarly, a new unit type "transient" is used
3343 for unit files created using the runtime API. "systemctl enable" will
3344 refuse to operate on such files.
3346 * A new command "systemctl revert" has been added that may be used to
3347 revert to the vendor version of a unit file, in case local changes
3348 have been made by adding drop-ins or overriding the unit file.
3350 * "machinectl clean" gained a new verb to automatically remove all or
3351 just hidden container images.
3353 * systemd-tmpfiles gained support for a new line type "e" for emptying
3354 directories, if they exist, without creating them if they don't.
3356 * systemd-nspawn gained support for automatically patching the UID/GIDs
3357 of the owners and the ACLs of all files and directories in a
3358 container tree to match the UID/GID user namespacing range selected
3359 for the container invocation. This mode is enabled via the new
3360 --private-users-chown switch. It also gained support for
3361 automatically choosing a free, previously unused UID/GID range when
3362 starting a container, via the new --private-users=pick setting (which
3363 implies --private-users-chown). Together, these options for the first
3364 time make user namespacing for nspawn containers fully automatic and
3365 thus deployable. The systemd-nspawn@.service template unit file has
3366 been changed to use this functionality by default.
3368 * systemd-nspawn gained a new --network-zone= switch, that allows
3369 creating ad-hoc virtual Ethernet links between multiple containers,
3370 that only exist as long as at least one container referencing them is
3371 running. This allows easy connecting of multiple containers with a
3372 common link that implements an Ethernet broadcast domain. Each of
3373 these network "zones" may be named relatively freely by the user, and
3374 may be referenced by any number of containers, but each container may
3375 only reference one of these "zones". On the lower level, this is
3376 implemented by an automatically managed bridge network interface for
3377 each zone, that is created when the first container referencing its
3378 zone is created and removed when the last one referencing its zone
3381 * The default start timeout may now be configured on the kernel command
3382 line via systemd.default_timeout_start_sec=. It was already
3383 configurable via the DefaultTimeoutStartSec= option in
3384 /etc/systemd/system.conf.
3386 * Socket units gained a new TriggerLimitIntervalSec= and
3387 TriggerLimitBurst= setting to configure a limit on the activation
3388 rate of the socket unit.
3390 * The LimitNICE= setting now optionally takes normal UNIX nice values
3391 in addition to the raw integer limit value. If the specified
3392 parameter is prefixed with "+" or "-" and is in the range -20..19 the
3393 value is understood as UNIX nice value. If not prefixed like this it
3394 is understood as raw RLIMIT_NICE limit.
3396 * Note that the effect of the PrivateDevices= unit file setting changed
3397 slightly with this release: the per-device /dev file system will be
3398 mounted read-only from this version on, and will have "noexec"
3399 set. This (minor) change of behavior might cause some (exceptional)
3400 legacy software to break, when PrivateDevices=yes is set for its
3401 service. Please leave PrivateDevices= off if you run into problems
3404 * systemd-bootchart has been split out to a separate repository:
3405 https://github.com/systemd/systemd-bootchart
3407 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
3408 merged into the kernel in its current form.
3410 * The compatibility libraries libsystemd-daemon.so,
3411 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
3412 which have been deprecated since systemd-209 have been removed along
3413 with the corresponding pkg-config files. All symbols provided by
3414 those libraries are provided by libsystemd.so.
3416 * The Capabilities= unit file setting has been removed (it is ignored
3417 for backwards compatibility). AmbientCapabilities= and
3418 CapabilityBoundingSet= should be used instead.
3420 * A new special target has been added, initrd-root-device.target,
3421 which creates a synchronization point for dependencies of the root
3422 device in early userspace. Initramfs builders must ensure that this
3423 target is now included in early userspace.
3425 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
3426 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
3427 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
3428 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
3429 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
3430 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
3431 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
3432 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
3433 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
3434 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
3435 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
3436 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
3437 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
3438 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
3439 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
3440 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
3441 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
3442 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
3443 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
3444 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
3445 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
3446 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
3447 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
3448 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
3449 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
3452 — Fairfax, 2016-05-21
3456 * The systemd-resolved DNS resolver service has gained a substantial
3457 set of new features, most prominently it may now act as a DNSSEC
3458 validating stub resolver. DNSSEC mode is currently turned off by
3459 default, but is expected to be turned on by default in one of the
3460 next releases. For now, we invite everybody to test the DNSSEC logic
3461 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
3462 service also gained a full set of D-Bus interfaces, including calls
3463 to configure DNS and DNSSEC settings per link (for use by external
3464 network management software). systemd-resolved and systemd-networkd
3465 now distinguish between "search" and "routing" domains. The former
3466 are used to qualify single-label names, the latter are used purely
3467 for routing lookups within certain domains to specific links.
3468 resolved now also synthesizes RRs for all entries from /etc/hosts.
3470 * The systemd-resolve tool (which is a client utility for
3471 systemd-resolved) has been improved considerably and is now fully
3472 supported and documented. Hence it has moved from /usr/lib/systemd to
3475 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
3478 * The coredump collection logic has been reworked: when a coredump is
3479 collected it is now written to disk, compressed and processed
3480 (including stacktrace extraction) from a new instantiated service
3481 systemd-coredump@.service, instead of directly from the
3482 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
3483 processing large coredumps can take up a substantial amount of
3484 resources and time, and this previously happened entirely outside of
3485 systemd's service supervision. With the new logic the core_pattern
3486 hook only does minimal metadata collection before passing off control
3487 to the new instantiated service, which is configured with a time
3488 limit, a nice level and other settings to minimize negative impact on
3489 the rest of the system. Also note that the new logic will honour the
3490 RLIMIT_CORE setting of the crashed process, which now allows users
3491 and processes to turn off coredumping for their processes by setting
3494 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
3495 and all forked processes by default. Previously, PID 1 would leave
3496 the setting at "0" for all processes, as set by the kernel. Note that
3497 the resource limit traditionally has no effect on the generated
3498 coredumps on the system if the /proc/sys/kernel/core_pattern hook
3499 logic is used. Since the limit is now honoured (see above) its
3500 default has been changed so that the coredumping logic is enabled by
3501 default for all processes, while allowing specific opt-out.
3503 * When the stacktrace is extracted from processes of system users, this
3504 is now done as "systemd-coredump" user, in order to sandbox this
3505 potentially security sensitive parsing operation. (Note that when
3506 processing coredumps of normal users this is done under the user ID
3507 of process that crashed, as before.) Packagers should take notice
3508 that it is now necessary to create the "systemd-coredump" system user
3509 and group at package installation time.
3511 * The systemd-activate socket activation testing tool gained support
3512 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
3513 and --seqpacket switches. It also has been extended to support both
3514 new-style and inetd-style file descriptor passing. Use the new
3515 --inetd switch to request inetd-style file descriptor passing.
3517 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
3518 variable, which takes a boolean value. If set to false, ANSI color
3519 output is disabled in the tools even when run on a terminal that
3522 * The VXLAN support in networkd now supports two new settings
3523 DestinationPort= and PortRange=.
3525 * A new systemd.machine_id= kernel command line switch has been added,
3526 that may be used to set the machine ID in /etc/machine-id if it is
3527 not initialized yet. This command line option has no effect if the
3528 file is already initialized.
3530 * systemd-nspawn gained a new --as-pid2 switch that invokes any
3531 specified command line as PID 2 rather than PID 1 in the
3532 container. In this mode PID 1 is a minimal stub init process that
3533 implements the special POSIX and Linux semantics of PID 1 regarding
3534 signal and child process management. Note that this stub init process
3535 is implemented in nspawn itself and requires no support from the
3536 container image. This new logic is useful to support running
3537 arbitrary commands in the container, as normal processes are
3538 generally not prepared to run as PID 1.
3540 * systemd-nspawn gained a new --chdir= switch for setting the current
3541 working directory for the process started in the container.
3543 * "journalctl /dev/sda" will now output all kernel log messages for
3544 specified device from the current boot, in addition to all devices
3545 that are parents of it. This should make log output about devices
3546 pretty useful, as long as kernel drivers attach enough metadata to
3547 the log messages. (The usual SATA drivers do.)
3549 * The sd-journal API gained two new calls
3550 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
3551 that report whether log data from /run or /var has been found.
3553 * journalctl gained a new switch "--fields" that prints all journal
3554 record field names currently in use in the journal. This is backed
3555 by two new sd-journal API calls sd_journal_enumerate_fields() and
3556 sd_journal_restart_fields().
3558 * Most configurable timeouts in systemd now expect an argument of
3559 "infinity" to turn them off, instead of "0" as before. The semantics
3560 from now on is that a timeout of "0" means "now", and "infinity"
3561 means "never". To maintain backwards compatibility, "0" continues to
3562 turn off previously existing timeout settings.
3564 * "systemctl reload-or-try-restart" has been renamed to "systemctl
3565 try-reload-or-restart" to clarify what it actually does: the "try"
3566 logic applies to both reloading and restarting, not just restarting.
3567 The old name continues to be accepted for compatibility.
3569 * On boot-up, when PID 1 detects that the system clock is behind the
3570 release date of the systemd version in use, the clock is now set
3571 to the latter. Previously, this was already done in timesyncd, in order
3572 to avoid running with clocks set to the various clock epochs such as
3573 1902, 1938 or 1970. With this change the logic is now done in PID 1
3574 in addition to timesyncd during early boot-up, so that it is enforced
3575 before the first process is spawned by systemd. Note that the logic
3576 in timesyncd remains, as it is more comprehensive and ensures
3577 clock monotonicity by maintaining a persistent timestamp file in
3578 /var. Since /var is generally not available in earliest boot or the
3579 initrd, this part of the logic remains in timesyncd, and is not done
3582 * Support for tweaking details in net_cls.class_id through the
3583 NetClass= configuration directive has been removed, as the kernel
3584 people have decided to deprecate that controller in cgroup v2.
3585 Userspace tools such as nftables are moving over to setting rules
3586 that are specific to the full cgroup path of a task, which obsoletes
3587 these controllers anyway. The NetClass= directive is kept around for
3588 legacy compatibility reasons. For a more in-depth description of the
3589 kernel change, please refer to the respective upstream commit:
3591 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
3593 * A new service setting RuntimeMaxSec= has been added that may be used
3594 to specify a maximum runtime for a service. If the timeout is hit, the
3595 service is terminated and put into a failure state.
3597 * A new service setting AmbientCapabilities= has been added. It allows
3598 configuration of additional Linux process capabilities that are
3599 passed to the activated processes. This is only available on very
3602 * The process resource limit settings in service units may now be used
3603 to configure hard and soft limits individually.
3605 * The various libsystemd APIs such as sd-bus or sd-event now publicly
3606 expose support for gcc's __attribute__((cleanup())) C extension.
3607 Specifically, for many object destructor functions alternative
3608 versions have been added that have names suffixed with "p" and take a
3609 pointer to a pointer to the object to destroy, instead of just a
3610 pointer to the object itself. This is useful because these destructor
3611 functions may be used directly as parameters to the cleanup
3612 construct. Internally, systemd has been a heavy user of this GCC
3613 extension for a long time, and with this change similar support is
3614 now available to consumers of the library outside of systemd. Note
3615 that by using this extension in your sources compatibility with old
3616 and strictly ANSI compatible C compilers is lost. However, all gcc or
3617 LLVM versions of recent years support this extension.
3619 * Timer units gained support for a new setting RandomizedDelaySec= that
3620 allows configuring some additional randomized delay to the configured
3621 time. This is useful to spread out timer events to avoid load peaks in
3622 clusters or larger setups.
3624 * Calendar time specifications now support sub-second accuracy.
3626 * Socket units now support listening on SCTP and UDP-lite protocol
3629 * The sd-event API now comes with a full set of man pages.
3631 * Older versions of systemd contained experimental support for
3632 compressing journal files and coredumps with the LZ4 compressor that
3633 was not compatible with the lz4 binary (due to API limitations of the
3634 lz4 library). This support has been removed; only support for files
3635 compatible with the lz4 binary remains. This LZ4 logic is now
3636 officially supported and no longer considered experimental.
3638 * The dkr image import logic has been removed again from importd. dkr's
3639 micro-services focus doesn't fit into the machine image focus of
3640 importd, and quickly got out of date with the upstream dkr API.
3642 * Creation of the /run/lock/lockdev/ directory was dropped from
3643 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
3644 been available for many years. If you still need this, you need to
3645 create your own tmpfiles.d config file with:
3647 d /run/lock/lockdev 0775 root lock -
3649 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
3650 and RebootArgument= have been moved from the [Service] section of
3651 unit files to [Unit], and they are now supported on all unit types,
3652 not just service units. Of course, systemd will continue to
3653 understand these settings also at the old location, in order to
3654 maintain compatibility.
3656 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
3657 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
3658 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
3659 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
3660 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
3661 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
3662 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
3663 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
3664 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
3665 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
3666 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
3667 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
3668 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
3669 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
3670 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
3671 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
3672 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
3673 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
3674 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
3676 — Berlin, 2016-02-11
3680 * A number of properties previously only settable in unit
3681 files are now also available as properties to set when
3682 creating transient units programmatically via the bus, as it
3683 is exposed with systemd-run's --property=
3684 setting. Specifically, these are: SyslogIdentifier=,
3685 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
3686 EnvironmentFile=, ReadWriteDirectories=,
3687 ReadOnlyDirectories=, InaccessibleDirectories=,
3688 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
3690 * When creating transient services via the bus API it is now
3691 possible to pass in a set of file descriptors to use as
3692 STDIN/STDOUT/STDERR for the invoked process.
3694 * Slice units may now be created transiently via the bus APIs,
3695 similar to the way service and scope units may already be
3696 created transiently.
3698 * Wherever systemd expects a calendar timestamp specification
3699 (like in journalctl's --since= and --until= switches) UTC
3700 timestamps are now supported. Timestamps suffixed with "UTC"
3701 are now considered to be in Universal Time Coordinated
3702 instead of the local timezone. Also, timestamps may now
3703 optionally be specified with sub-second accuracy. Both of
3704 these additions also apply to recurring calendar event
3705 specification, such as OnCalendar= in timer units.
3707 * journalctl gained a new "--sync" switch that asks the
3708 journal daemon to write all so far unwritten log messages to
3709 disk and sync the files, before returning.
3711 * systemd-tmpfiles learned two new line types "q" and "Q" that
3712 operate like "v", but also set up a basic btrfs quota
3713 hierarchy when used on a btrfs file system with quota
3716 * tmpfiles' "v", "q" and "Q" will now create a plain directory
3717 instead of a subvolume (even on a btrfs file system) if the
3718 root directory is a plain directory, and not a
3719 subvolume. This should simplify things with certain chroot()
3720 environments which are not aware of the concept of btrfs
3723 * systemd-detect-virt gained a new --chroot switch to detect
3724 whether execution takes place in a chroot() environment.
3726 * CPUAffinity= now takes CPU index ranges in addition to
3729 * The various memory-related resource limit settings (such as
3730 LimitAS=) now understand the usual K, M, G, ... suffixes to
3731 the base of 1024 (IEC). Similar, the time-related resource
3732 limit settings understand the usual min, h, day, ...
3735 * There's a new system.conf setting DefaultTasksMax= to
3736 control the default TasksMax= setting for services and
3737 scopes running on the system. (TasksMax= is the primary
3738 setting that exposes the "pids" cgroup controller on systemd
3739 and was introduced in the previous systemd release.) The
3740 setting now defaults to 512, which means services that are
3741 not explicitly configured otherwise will only be able to
3742 create 512 processes or threads at maximum, from this
3743 version on. Note that this means that thread- or
3744 process-heavy services might need to be reconfigured to set
3745 TasksMax= to a higher value. It is sufficient to set
3746 TasksMax= in these specific unit files to a higher value, or
3747 even "infinity". Similar, there's now a logind.conf setting
3748 UserTasksMax= that defaults to 4096 and limits the total
3749 number of processes or tasks each user may own
3750 concurrently. nspawn containers also have the TasksMax=
3751 value set by default now, to 8192. Note that all of this
3752 only has an effect if the "pids" cgroup controller is
3753 enabled in the kernel. The general benefit of these changes
3754 should be a more robust and safer system, that provides a
3755 certain amount of per-service fork() bomb protection.
3757 * systemd-nspawn gained the new --network-veth-extra= switch
3758 to define additional and arbitrarily-named virtual Ethernet
3759 links between the host and the container.
3761 * A new service execution setting PassEnvironment= has been
3762 added that allows importing select environment variables
3763 from PID1's environment block into the environment block of
3766 * Timer units gained support for a new RemainAfterElapse=
3767 setting which takes a boolean argument. It defaults to on,
3768 exposing behaviour unchanged to previous releases. If set to
3769 off, timer units are unloaded after they elapsed if they
3770 cannot elapse again. This is particularly useful for
3771 transient timer units, which shall not stay around longer
3772 than until they first elapse.
3774 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
3775 default now (the kernel default is 16). This is beneficial
3776 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
3777 allows substantially larger numbers of queued
3778 datagrams. This should increase the capability of systemd to
3779 parallelize boot-up, as logging and sd_notify() are unlikely
3780 to stall execution anymore. If you need to change the value
3781 from the new defaults, use the usual sysctl.d/ snippets.
3783 * The compression framing format used by the journal or
3784 coredump processing has changed to be in line with what the
3785 official LZ4 tools generate. LZ4 compression support in
3786 systemd was considered unsupported previously, as the format
3787 was not compatible with the normal tools. With this release
3788 this has changed now, and it is hence safe for downstream
3789 distributions to turn it on. While not compressing as well
3790 as the XZ, LZ4 is substantially faster, which makes
3791 it a good default choice for the compression logic in the
3792 journal and in coredump handling.
3794 * Any reference to /etc/mtab has been dropped from
3795 systemd. The file has been obsolete since a while, but
3796 systemd refused to work on systems where it was incorrectly
3797 set up (it should be a symlink or non-existent). Please make
3798 sure to update to util-linux 2.27.1 or newer in conjunction
3799 with this systemd release, which also drops any reference to
3800 /etc/mtab. If you maintain a distribution make sure that no
3801 software you package still references it, as this is a
3802 likely source of bugs. There's also a glibc bug pending,
3803 asking for removal of any reference to this obsolete file:
3805 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
3807 Note that only util-linux versions built with
3808 --enable-libmount-force-mountinfo are supported.
3810 * Support for the ".snapshot" unit type has been removed. This
3811 feature turned out to be little useful and little used, and
3812 has now been removed from the core and from systemctl.
3814 * The dependency types RequiresOverridable= and
3815 RequisiteOverridable= have been removed from systemd. They
3816 have been used only very sparingly to our knowledge and
3817 other options that provide a similar effect (such as
3818 systemctl --mode=ignore-dependencies) are much more useful
3819 and commonly used. Moreover, they were only half-way
3820 implemented as the option to control behaviour regarding
3821 these dependencies was never added to systemctl. By removing
3822 these dependency types the execution engine becomes a bit
3823 simpler. Unit files that use these dependencies should be
3824 changed to use the non-Overridable dependency types
3825 instead. In fact, when parsing unit files with these
3826 options, that's what systemd will automatically convert them
3827 too, but it will also warn, asking users to fix the unit
3828 files accordingly. Removal of these dependency types should
3829 only affect a negligible number of unit files in the wild.
3831 * Behaviour of networkd's IPForward= option changed
3832 (again). It will no longer maintain a per-interface setting,
3833 but propagate one way from interfaces where this is enabled
3834 to the global kernel setting. The global setting will be
3835 enabled when requested by a network that is set up, but
3836 never be disabled again. This change was made to make sure
3837 IPv4 and IPv6 behaviour regarding packet forwarding is
3838 similar (as the Linux IPv6 stack does not support
3839 per-interface control of this setting) and to minimize
3842 * In unit files the behaviour of %u, %U, %h, %s has
3843 changed. These specifiers will now unconditionally resolve
3844 to the various user database fields of the user that the
3845 systemd instance is running as, instead of the user
3846 configured in the specific unit via User=. Note that this
3847 effectively doesn't change much, as resolving of these
3848 specifiers was already turned off in the --system instance
3849 of systemd, as we cannot do NSS lookups from PID 1. In the
3850 --user instance of systemd these specifiers where correctly
3851 resolved, but hardly made any sense, since the user instance
3852 lacks privileges to do user switches anyway, and User= is
3853 hence useless. Moreover, even in the --user instance of
3854 systemd behaviour was awkward as it would only take settings
3855 from User= assignment placed before the specifier into
3856 account. In order to unify and simplify the logic around
3857 this the specifiers will now always resolve to the
3858 credentials of the user invoking the manager (which in case
3859 of PID 1 is the root user).
3861 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
3862 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
3863 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
3864 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
3865 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
3866 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
3867 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
3868 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
3869 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
3870 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
3871 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
3872 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
3873 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
3874 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
3877 — Berlin, 2015-11-18
3881 * systemd now depends on util-linux v2.27. More specifically,
3882 the newly added mount monitor feature in libmount now
3883 replaces systemd's former own implementation.
3885 * libmount mandates /etc/mtab not to be regular file, and
3886 systemd now enforces this condition at early boot.
3887 /etc/mtab has been deprecated and warned about for a very
3888 long time, so systems running systemd should already have
3889 stopped having this file around as anything else than a
3890 symlink to /proc/self/mounts.
3892 * Support for the "pids" cgroup controller has been added. It
3893 allows accounting the number of tasks in a cgroup and
3894 enforcing limits on it. This adds two new setting
3895 TasksAccounting= and TasksMax= to each unit, as well as a
3896 global option DefaultTasksAccounting=.
3898 * Support for the "net_cls" cgroup controller has been added.
3899 It allows assigning a net class ID to each task in the
3900 cgroup, which can then be used in firewall rules and traffic
3901 shaping configurations. Note that the kernel netfilter net
3902 class code does not currently work reliably for ingress
3903 packets on unestablished sockets.
3905 This adds a new config directive called NetClass= to CGroup
3906 enabled units. Allowed values are positive numbers for fixed
3907 assignments and "auto" for picking a free value
3910 * 'systemctl is-system-running' now returns 'offline' if the
3911 system is not booted with systemd. This command can now be
3912 used as a substitute for 'systemd-notify --booted'.
3914 * Watchdog timeouts have been increased to 3 minutes for all
3915 in-tree service files. Apparently, disk IO issues are more
3916 frequent than we hoped, and user reported >1 minute waiting
3919 * 'machine-id-commit' functionality has been merged into
3920 'machine-id-setup --commit'. The separate binary has been
3923 * The WorkingDirectory= directive in unit files may now be set
3924 to the special value '~'. In this case, the working
3925 directory is set to the home directory of the user
3926 configured in User=.
3928 * "machinectl shell" will now open the shell in the home
3929 directory of the selected user by default.
3931 * The CrashChVT= configuration file setting is renamed to
3932 CrashChangeVT=, following our usual logic of not
3933 abbreviating unnecessarily. The old directive is still
3934 supported for compat reasons. Also, this directive now takes
3935 an integer value between 1 and 63, or a boolean value. The
3936 formerly supported '-1' value for disabling stays around for
3939 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
3940 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
3941 RootDirectory= properties can now be set for transient
3944 * The systemd-analyze tool gained a new "set-log-target" verb
3945 to change the logging target the system manager logs to
3946 dynamically during runtime. This is similar to how
3947 "systemd-analyze set-log-level" already changes the log
3950 * In nspawn /sys is now mounted as tmpfs, with only a selected
3951 set of subdirectories mounted in from the real sysfs. This
3952 enhances security slightly, and is useful for ensuring user
3953 namespaces work correctly.
3955 * Support for USB FunctionFS activation has been added. This
3956 allows implementation of USB gadget services that are
3957 activated as soon as they are requested, so that they don't
3958 have to run continuously, similar to classic socket
3961 * The "systemctl exit" command now optionally takes an
3962 additional parameter that sets the exit code to return from
3963 the systemd manager when exiting. This is only relevant when
3964 running the systemd user instance, or when running the
3965 system instance in a container.
3967 * sd-bus gained the new API calls sd_bus_path_encode_many()
3968 and sd_bus_path_decode_many() that allow easy encoding and
3969 decoding of multiple identifier strings inside a D-Bus
3970 object path. Another new call sd_bus_default_flush_close()
3971 has been added to flush and close per-thread default
3974 * systemd-cgtop gained support for a -M/--machine= switch to
3975 show the control groups within a certain container only.
3977 * "systemctl kill" gained support for an optional --fail
3978 switch. If specified the requested operation will fail of no
3979 processes have been killed, because the unit had no
3980 processes attached, or similar.
3982 * A new systemd.crash_reboot=1 kernel command line option has
3983 been added that triggers a reboot after crashing. This can
3984 also be set through CrashReboot= in systemd.conf.
3986 * The RuntimeDirectory= setting now understands unit
3987 specifiers like %i or %f.
3989 * A new (still internal) library API sd-ipv4acd has been added,
3990 that implements address conflict detection for IPv4. It's
3991 based on code from sd-ipv4ll, and will be useful for
3992 detecting DHCP address conflicts.
3994 * File descriptors passed during socket activation may now be
3995 named. A new API sd_listen_fds_with_names() is added to
3996 access the names. The default names may be overridden,
3997 either in the .socket file using the FileDescriptorName=
3998 parameter, or by passing FDNAME= when storing the file
3999 descriptors using sd_notify().
4001 * systemd-networkd gained support for:
4003 - Setting the IPv6 Router Advertisement settings via
4004 IPv6AcceptRouterAdvertisements= in .network files.
4006 - Configuring the HelloTimeSec=, MaxAgeSec= and
4007 ForwardDelaySec= bridge parameters in .netdev files.
4009 - Configuring PreferredSource= for static routes in
4012 * The "ask-password" framework used to query for LUKS harddisk
4013 passwords or SSL passwords during boot gained support for
4014 caching passwords in the kernel keyring, if it is
4015 available. This makes sure that the user only has to type in
4016 a passphrase once if there are multiple objects to unlock
4017 with the same one. Previously, such password caching was
4018 available only when Plymouth was used; this moves the
4019 caching logic into the systemd codebase itself. The
4020 "systemd-ask-password" utility gained a new --keyname=
4021 switch to control which kernel keyring key to use for
4022 caching a password in. This functionality is also useful for
4023 enabling display managers such as gdm to automatically
4024 unlock the user's GNOME keyring if its passphrase, the
4025 user's password and the harddisk password are the same, if
4026 gdm-autologin is used.
4028 * When downloading tar or raw images using "machinectl
4029 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
4030 file is now also downloaded, if it is available and stored
4031 next to the image file.
4033 * Units of type ".socket" gained a new boolean setting
4034 Writable= which is only useful in conjunction with
4035 ListenSpecial=. If true, enables opening the specified
4036 special file in O_RDWR mode rather than O_RDONLY mode.
4038 * systemd-rfkill has been reworked to become a singleton
4039 service that is activated through /dev/rfkill on each rfkill
4040 state change and saves the settings to disk. This way,
4041 systemd-rfkill is now compatible with devices that exist
4042 only intermittendly, and even restores state if the previous
4043 system shutdown was abrupt rather than clean.
4045 * The journal daemon gained support for vacuuming old journal
4046 files controlled by the number of files that shall remain,
4047 in addition to the already existing control by size and by
4048 date. This is useful as journal interleaving performance
4049 degrades with too many separate journal files, and allows
4050 putting an effective limit on them. The new setting defaults
4051 to 100, but this may be changed by setting SystemMaxFiles=
4052 and RuntimeMaxFiles= in journald.conf. Also, the
4053 "journalctl" tool gained the new --vacuum-files= switch to
4054 manually vacuum journal files to leave only the specified
4055 number of files in place.
4057 * udev will now create /dev/disk/by-path links for ATA devices
4058 on kernels where that is supported.
4060 * Galician, Serbian, Turkish and Korean translations were added.
4062 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
4063 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
4064 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
4065 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
4066 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
4067 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
4068 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
4069 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
4070 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
4071 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
4072 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
4073 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
4074 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
4075 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
4076 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
4077 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
4078 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
4079 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
4081 — Berlin, 2015-10-07
4085 * The DHCP implementation of systemd-networkd gained a set of
4088 - The DHCP server now supports emitting DNS and NTP
4089 information. It may be enabled and configured via
4090 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
4091 and NTP information is enabled, but no servers are
4092 configured, the corresponding uplink information (if there
4093 is any) is propagated.
4095 - Server and client now support transmission and reception
4096 of timezone information. It can be configured via the
4097 newly introduced network options UseTimezone=,
4098 EmitTimezone=, and Timezone=. Transmission of timezone
4099 information is enabled between host and containers by
4100 default now: the container will change its local timezone
4101 to what the host has set.
4103 - Lease timeouts can now be configured via
4104 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
4106 - The DHCP server improved on the stability of
4107 leases. Clients are more likely to get the same lease
4108 information back, even if the server loses state.
4110 - The DHCP server supports two new configuration options to
4111 control the lease address pool metrics, PoolOffset= and
4114 * The encapsulation limit of tunnels in systemd-networkd may
4115 now be configured via 'EncapsulationLimit='. It allows
4116 modifying the maximum additional levels of encapsulation
4117 that are permitted to be prepended to a packet.
4119 * systemd now supports the concept of user buses replacing
4120 session buses, if used with dbus-1.10 (and enabled via dbus
4121 --enable-user-session). It previously only supported this on
4122 kdbus-enabled systems, and this release expands this to
4123 'dbus-daemon' systems.
4125 * systemd-networkd now supports predictable interface names
4128 * systemd now optionally supports the new Linux kernel
4129 "unified" control group hierarchy. If enabled via the kernel
4130 command-line option 'systemd.unified_cgroup_hierarchy=1',
4131 systemd will try to mount the unified cgroup hierarchy
4132 directly on /sys/fs/cgroup. If not enabled, or not
4133 available, systemd will fall back to the legacy cgroup
4134 hierarchy setup, as before. Host system and containers can
4135 mix and match legacy and unified hierarchies as they
4136 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
4137 environment variable to individually select the hierarchy to
4138 use for executed containers. By default, nspawn will use the
4139 unified hierarchy for the containers if the host uses the
4140 unified hierarchy, and the legacy hierarchy otherwise.
4141 Please note that at this point the unified hierarchy is an
4142 experimental kernel feature and is likely to change in one
4143 of the next kernel releases. Therefore, it should not be
4144 enabled by default in downstream distributions yet. The
4145 minimum required kernel version for the unified hierarchy to
4146 work is 4.2. Note that when the unified hierarchy is used
4147 for the first time delegated access to controllers is
4148 safe. Because of this systemd-nspawn containers will get
4149 access to controllers now, as will systemd user
4150 sessions. This means containers and user sessions may now
4151 manage their own resources, partitioning up what the system
4154 * A new special scope unit "init.scope" has been introduced
4155 that encapsulates PID 1 of the system. It may be used to
4156 determine resource usage and enforce resource limits on PID
4157 1 itself. PID 1 hence moved out of the root of the control
4160 * The cgtop tool gained support for filtering out kernel
4161 threads when counting tasks in a control group. Also, the
4162 count of processes is now recursively summed up by
4163 default. Two options -k and --recursive= have been added to
4164 revert to old behaviour. The tool has also been updated to
4165 work correctly in containers now.
4167 * systemd-nspawn's --bind= and --bind-ro= options have been
4168 extended to allow creation of non-recursive bind mounts.
4170 * libsystemd gained two new calls sd_pid_get_cgroup() and
4171 sd_peer_get_cgroup() which return the control group path of
4172 a process or peer of a connected AF_UNIX socket. This
4173 function call is particularly useful when implementing
4174 delegated subtrees support in the control group hierarchy.
4176 * The "sd-event" event loop API of libsystemd now supports
4177 correct dequeuing of real-time signals, without losing
4180 * When systemd requests a polkit decision when managing units it
4181 will now add additional fields to the request, including unit
4182 name and desired operation. This enables more powerful polkit
4183 policies, that make decisions depending on these parameters.
4185 * nspawn learnt support for .nspawn settings files, that may
4186 accompany the image files or directories of containers, and
4187 may contain additional settings for the container. This is
4188 an alternative to configuring container parameters via the
4189 nspawn command line.
4191 Contributions from: Cristian Rodríguez, Daniel Mack, David
4192 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
4193 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
4194 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
4195 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
4196 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
4197 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
4198 Andersen, Tom Gundersen, Torstein Husebø
4200 — Berlin, 2015-09-08
4204 * machinectl gained a new verb 'shell' which opens a fresh
4205 shell on the target container or the host. It is similar to
4206 the existing 'login' command of machinectl, but spawns the
4207 shell directly without prompting for username or
4208 password. The pseudo machine '.host' now refers to the local
4209 host and is used by default. Hence, 'machinectl shell' can
4210 be used as replacement for 'su -' which spawns a session as
4211 a fresh systemd unit in a way that is fully isolated from
4212 the originating session.
4214 * systemd-networkd learned to cope with private-zone DHCP
4215 options and allows other programs to query the values.
4217 * SELinux access control when enabling/disabling units is no
4218 longer enforced with this release. The previous implementation
4219 was incorrect, and a new corrected implementation is not yet
4220 available. As unit file operations are still protected via
4221 polkit and D-Bus policy this is not a security problem. Yet,
4222 distributions which care about optimal SELinux support should
4223 probably not stabilize on this release.
4225 * sd-bus gained support for matches of type "arg0has=", that
4226 test for membership of strings in string arrays sent in bus
4229 * systemd-resolved now dumps the contents of its DNS and LLMNR
4230 caches to the logs on reception of the SIGUSR1 signal. This
4231 is useful to debug DNS behaviour.
4233 * The coredumpctl tool gained a new --directory= option to
4234 operate on journal files in a specific directory.
4236 * "systemctl reboot" and related commands gained a new
4237 "--message=" option which may be used to set a free-text
4238 wall message when shutting down or rebooting the
4239 system. This message is also logged, which is useful for
4240 figuring out the reason for a reboot or shutdown a
4243 * The "systemd-resolve-host" tool's -i switch now takes
4244 network interface numbers as alternative to interface names.
4246 * A new unit file setting for services has been introduced:
4247 UtmpMode= allows configuration of how precisely systemd
4248 handles utmp and wtmp entries for the service if this is
4249 enabled. This allows writing services that appear similar to
4250 user sessions in the output of the "w", "who", "last" and
4253 * systemd-resolved will now locally synthesize DNS resource
4254 records for the "localhost" and "gateway" domains as well as
4255 the local hostname. This should ensure that clients querying
4256 RRs via resolved will get similar results as those going via
4257 NSS, if nss-myhostname is enabled.
4259 Contributions from: Alastair Hughes, Alex Crawford, Daniel
4260 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
4261 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
4262 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
4263 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
4264 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
4265 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
4266 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
4267 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
4268 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
4269 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
4270 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4272 — Berlin, 2015-08-27
4276 * The systemd-efi-boot-generator functionality was merged into
4277 systemd-gpt-auto-generator.
4279 * systemd-networkd now supports Group Policy for vxlan
4280 devices. It can be enabled via the new boolean configuration
4281 option called 'GroupPolicyExtension='.
4283 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
4284 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
4285 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
4287 — Berlin, 2015-07-31
4291 * The python-systemd code has been removed from the systemd repository.
4292 A new repository has been created which accommodates the code from
4293 now on, and we kindly ask distributions to create a separate package
4294 for this: https://github.com/systemd/python-systemd
4296 * The systemd daemon will now reload its main configuration
4297 (/etc/systemd/system.conf) on daemon-reload.
4299 * sd-dhcp now exposes vendor specific extensions via
4300 sd_dhcp_lease_get_vendor_specific().
4302 * systemd-networkd gained a number of new configuration options.
4304 - A new boolean configuration option for TAP devices called
4305 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
4306 device, thus allowing to send and receive GSO packets.
4308 - A new tunnel configuration option called 'CopyDSCP='.
4309 If enabled, the DSCP field of ip6 tunnels is copied into the
4310 decapsulated packet.
4312 - A set of boolean bridge configuration options were added.
4313 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
4314 and 'UnicastFlood=' are now parsed by networkd and applied to the
4315 respective bridge link device via the respective IFLA_BRPORT_*
4318 - A new string configuration option to override the hostname sent
4319 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
4320 is true, networkd will use the configured hostname instead of the
4321 system hostname when sending DHCP requests.
4323 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
4324 networkd will configure the IPv6 flow-label of the tunnel device
4325 according to RFC2460.
4327 - The 'macvtap' virtual network devices are now supported, similar to
4328 the already supported 'macvlan' devices.
4330 * systemd-resolved now implements RFC5452 to improve resilience against
4331 cache poisoning. Additionally, source port randomization is enabled
4332 by default to further protect against DNS spoofing attacks.
4334 * nss-mymachines now supports translating UIDs and GIDs of running
4335 containers with user-namespaces enabled. If a container 'foo'
4336 translates a host uid 'UID' to the container uid 'TUID', then
4337 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
4338 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
4339 mapped as 'vg-foo-TGID'.
4341 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
4342 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
4343 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
4344 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
4345 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
4346 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
4347 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
4348 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
4349 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
4350 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
4352 — Berlin, 2015-07-29
4356 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
4357 There are no known issues with current sysfs, and udev does not need
4358 or should be used to work around such bugs.
4360 * udev does no longer enable USB HID power management. Several reports
4361 indicate, that some devices cannot handle that setting.
4363 * The udev accelerometer helper was removed. The functionality
4364 is now fully included in iio-sensor-proxy. But this means,
4365 older iio-sensor-proxy versions will no longer provide
4366 accelerometer/orientation data with this systemd version.
4367 Please upgrade iio-sensor-proxy to version 1.0.
4369 * networkd gained a new configuration option IPv6PrivacyExtensions=
4370 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
4371 for Stateless Address") on selected networks.
4373 * For the sake of fewer build-time dependencies and less code in the
4374 main repository, the python bindings are about to be removed in the
4375 next release. A new repository has been created which accommodates
4376 the code from now on, and we kindly ask distributions to create a
4377 separate package for this. The removal will take place in v223.
4379 https://github.com/systemd/python-systemd
4381 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
4382 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
4383 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
4384 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
4385 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
4386 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
4387 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
4388 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
4389 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4390 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4392 — Berlin, 2015-07-07
4396 * The sd-bus.h and sd-event.h APIs have now been declared
4397 stable and have been added to the official interface of
4398 libsystemd.so. sd-bus implements an alternative D-Bus client
4399 library, that is relatively easy to use, very efficient and
4400 supports both classic D-Bus as well as kdbus as transport
4401 backend. sd-event is a generic event loop abstraction that
4402 is built around Linux epoll, but adds features such as event
4403 prioritization or efficient timer handling. Both APIs are good
4404 choices for C programs looking for a bus and/or event loop
4405 implementation that is minimal and does not have to be
4406 portable to other kernels.
4408 * kdbus support is no longer compile-time optional. It is now
4409 always built-in. However, it can still be disabled at
4410 runtime using the kdbus=0 kernel command line setting, and
4411 that setting may be changed to default to off, by specifying
4412 --disable-kdbus at build-time. Note though that the kernel
4413 command line setting has no effect if the kdbus.ko kernel
4414 module is not installed, in which case kdbus is (obviously)
4415 also disabled. We encourage all downstream distributions to
4416 begin testing kdbus by adding it to the kernel images in the
4417 development distributions, and leaving kdbus support in
4420 * The minimal required util-linux version has been bumped to
4423 * Support for chkconfig (--enable-chkconfig) was removed in
4424 favor of calling an abstraction tool
4425 /lib/systemd/systemd-sysv-install. This needs to be
4426 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
4427 in README for details.
4429 * If there's a systemd unit and a SysV init script for the
4430 same service name, and the user executes "systemctl enable"
4431 for it (or a related call), then this will now enable both
4432 (or execute the related operation on both), not just the
4435 * The libudev API documentation has been converted from gtkdoc
4438 * gudev has been removed from the systemd tree, it is now an
4441 * The systemd-cgtop tool learnt a new --raw switch to generate
4442 "raw" (machine parsable) output.
4444 * networkd's IPForwarding= .network file setting learnt the
4445 new setting "kernel", which ensures that networkd does not
4446 change the IP forwarding sysctl from the default kernel
4449 * The systemd-logind bus API now exposes a new boolean
4450 property "Docked" that reports whether logind considers the
4451 system "docked", i.e. connected to a docking station or not.
4453 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
4454 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
4455 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
4456 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
4457 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
4458 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
4459 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
4460 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
4461 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
4462 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
4463 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
4464 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
4465 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
4466 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4467 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
4468 Fink, Zbigniew Jędrzejewski-Szmek
4470 — Berlin, 2015-06-19
4474 * The gudev library has been extracted into a separate repository
4475 available at: https://git.gnome.org/browse/libgudev/
4476 It is now managed as part of the Gnome project. Distributions
4477 are recommended to pass --disable-gudev to systemd and use
4478 gudev from the Gnome project instead. gudev is still included
4479 in systemd, for now. It will be removed soon, though. Please
4480 also see the announcement-thread on systemd-devel:
4481 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
4483 * systemd now exposes a CPUUsageNSec= property for each
4484 service unit on the bus, that contains the overall consumed
4485 CPU time of a service (the sum of what each process of the
4486 service consumed). This value is only available if
4487 CPUAccounting= is turned on for a service, and is then shown
4488 in the "systemctl status" output.
4490 * Support for configuring alternative mappings of the old SysV
4491 runlevels to systemd targets has been removed. They are now
4492 hardcoded in a way that runlevels 2, 3, 4 all map to
4493 multi-user.target and 5 to graphical.target (which
4494 previously was already the default behaviour).
4496 * The auto-mounter logic gained support for mount point
4497 expiry, using a new TimeoutIdleSec= setting in .automount
4498 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
4500 * The EFI System Partition (ESP) as mounted to /boot by
4501 systemd-efi-boot-generator will now be unmounted
4502 automatically after 2 minutes of not being used. This should
4503 minimize the risk of ESP corruptions.
4505 * New /etc/fstab options x-systemd.requires= and
4506 x-systemd.requires-mounts-for= are now supported to express
4507 additional dependencies for mounts. This is useful for
4508 journalling file systems that support external journal
4509 devices or overlay file systems that require underlying file
4510 systems to be mounted.
4512 * systemd does not support direct live-upgrades (via systemctl
4513 daemon-reexec) from versions older than v44 anymore. As no
4514 distribution we are aware of shipped such old versions in a
4515 stable release this should not be problematic.
4517 * When systemd forks off a new per-connection service instance
4518 it will now set the $REMOTE_ADDR environment variable to the
4519 remote IP address, and $REMOTE_PORT environment variable to
4520 the remote IP port. This behaviour is similar to the
4521 corresponding environment variables defined by CGI.
4523 * systemd-networkd gained support for uplink failure
4524 detection. The BindCarrier= option allows binding interface
4525 configuration dynamically to the link sense of other
4526 interfaces. This is useful to achieve behaviour like in
4529 * systemd-networkd gained support for configuring the DHCP
4530 client identifier to use when requesting leases.
4532 * systemd-networkd now has a per-network UseNTP= option to
4533 configure whether NTP server information acquired via DHCP
4534 is passed on to services like systemd-timesyncd.
4536 * systemd-networkd gained support for vti6 tunnels.
4538 * Note that systemd-networkd manages the sysctl variable
4539 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
4540 it is configured for since v219. The variable controls IP
4541 forwarding, and is a per-interface alternative to the global
4542 /proc/sys/net/ipv[46]/ip_forward. This setting is
4543 configurable in the IPForward= option, which defaults to
4544 "no". This means if networkd is used for an interface it is
4545 no longer sufficient to set the global sysctl option to turn
4546 on IP forwarding! Instead, the .network file option
4547 IPForward= needs to be turned on! Note that the
4548 implementation of this behaviour was broken in v219 and has
4551 * Many bonding and vxlan options are now configurable in
4554 * systemd-nspawn gained a new --property= setting to set unit
4555 properties for the container scope. This is useful for
4556 setting resource parameters (e.g. "CPUShares=500") on
4557 containers started from the command line.
4559 * systemd-nspawn gained a new --private-users= switch to make
4560 use of user namespacing available on recent Linux kernels.
4562 * systemd-nspawn may now be called as part of a shell pipeline
4563 in which case the pipes used for stdin and stdout are passed
4564 directly to the process invoked in the container, without
4565 indirection via a pseudo tty.
4567 * systemd-nspawn gained a new switch to control the UNIX
4568 signal to use when killing the init process of the container
4571 * systemd-nspawn gained a new --overlay= switch for mounting
4572 overlay file systems into the container using the new kernel
4575 * When a container image is imported via systemd-importd and
4576 the host file system is not btrfs, a loopback block device
4577 file is created in /var/lib/machines.raw with a btrfs file
4578 system inside. It is then mounted to /var/lib/machines to
4579 enable btrfs features for container management. The loopback
4580 file and btrfs file system is grown as needed when container
4581 images are imported via systemd-importd.
4583 * systemd-machined/systemd-importd gained support for btrfs
4584 quota, to enforce container disk space limits on disk. This
4585 is exposed in "machinectl set-limit".
4587 * systemd-importd now can import containers from local .tar,
4588 .raw and .qcow2 images, and export them to .tar and .raw. It
4589 can also import dkr v2 images now from the network (on top
4592 * systemd-importd gained support for verifying downloaded
4593 images with gpg2 (previously only gpg1 was supported).
4595 * systemd-machined, systemd-logind, systemd: most bus calls are
4596 now accessible to unprivileged processes via polkit. Also,
4597 systemd-logind will now allow users to kill their own sessions
4598 without further privileges or authorization.
4600 * systemd-shutdownd has been removed. This service was
4601 previously responsible for implementing scheduled shutdowns
4602 as exposed in /usr/bin/shutdown's time parameter. This
4603 functionality has now been moved into systemd-logind and is
4604 accessible via a bus interface.
4606 * "systemctl reboot" gained a new switch --firmware-setup that
4607 can be used to reboot into the EFI firmware setup, if that
4608 is available. systemd-logind now exposes an API on the bus
4609 to trigger such reboots, in case graphical desktop UIs want
4610 to cover this functionality.
4612 * "systemctl enable", "systemctl disable" and "systemctl mask"
4613 now support a new "--now" switch. If specified the units
4614 that are enabled will also be started, and the ones
4615 disabled/masked also stopped.
4617 * The Gummiboot EFI boot loader tool has been merged into
4618 systemd, and renamed to "systemd-boot". The bootctl tool has been
4619 updated to support systemd-boot.
4621 * An EFI kernel stub has been added that may be used to create
4622 kernel EFI binaries that contain not only the actual kernel,
4623 but also an initrd, boot splash, command line and OS release
4624 information. This combined binary can then be signed as a
4625 single image, so that the firmware can verify it all in one
4626 step. systemd-boot has special support for EFI binaries created
4627 like this and can extract OS release information from them
4628 and show them in the boot menu. This functionality is useful
4629 to implement cryptographically verified boot schemes.
4631 * Optional support has been added to systemd-fsck to pass
4632 fsck's progress report to an AF_UNIX socket in the file
4635 * udev will no longer create device symlinks for all block
4636 devices by default. A blacklist for excluding special block
4637 devices from this logic has been turned into a whitelist
4638 that requires picking block devices explicitly that require
4641 * A new (currently still internal) API sd-device.h has been
4642 added to libsystemd. This modernized API is supposed to
4643 replace libudev eventually. In fact, already much of libudev
4644 is now just a wrapper around sd-device.h.
4646 * A new hwdb database for storing metadata about pointing
4647 stick devices has been added.
4649 * systemd-tmpfiles gained support for setting file attributes
4650 similar to the "chattr" tool with new 'h' and 'H' lines.
4652 * systemd-journald will no longer unconditionally set the
4653 btrfs NOCOW flag on new journal files. This is instead done
4654 with tmpfiles snippet using the new 'h' line type. This
4655 allows easy disabling of this logic, by masking the
4656 journal-nocow.conf tmpfiles file.
4658 * systemd-journald will now translate audit message types to
4659 human readable identifiers when writing them to the
4660 journal. This should improve readability of audit messages.
4662 * The LUKS logic gained support for the offset= and skip=
4663 options in /etc/crypttab, as previously implemented by
4666 * /usr/lib/os-release gained a new optional field VARIANT= for
4667 distributions that support multiple variants (such as a
4668 desktop edition, a server edition, ...)
4670 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
4671 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
4672 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
4673 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
4674 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
4675 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
4676 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
4677 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
4678 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
4679 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
4680 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
4681 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
4682 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
4683 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
4684 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
4685 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
4686 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
4687 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
4688 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
4689 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
4690 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
4691 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
4692 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
4693 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
4694 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
4695 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
4696 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
4698 — Berlin, 2015-05-22
4702 * Introduce a new API "sd-hwdb.h" for querying the hardware
4703 metadata database. With this minimal interface one can query
4704 and enumerate the udev hwdb, decoupled from the old libudev
4705 library. libudev's interface for this is now only a wrapper
4706 around sd-hwdb. A new tool systemd-hwdb has been added to
4707 interface with and update the database.
4709 * When any of systemd's tools copies files (for example due to
4710 tmpfiles' C lines) a btrfs reflink will attempted first,
4711 before bytewise copying is done.
4713 * systemd-nspawn gained a new --ephemeral switch. When
4714 specified a btrfs snapshot is taken of the container's root
4715 directory, and immediately removed when the container
4716 terminates again. Thus, a container can be started whose
4717 changes never alter the container's root directory, and are
4718 lost on container termination. This switch can also be used
4719 for starting a container off the root file system of the
4720 host without affecting the host OS. This switch is only
4721 available on btrfs file systems.
4723 * systemd-nspawn gained a new --template= switch. It takes the
4724 path to a container tree to use as template for the tree
4725 specified via --directory=, should that directory be
4726 missing. This allows instantiating containers dynamically,
4727 on first run. This switch is only available on btrfs file
4730 * When a .mount unit refers to a mount point on which multiple
4731 mounts are stacked, and the .mount unit is stopped all of
4732 the stacked mount points will now be unmounted until no
4733 mount point remains.
4735 * systemd now has an explicit notion of supported and
4736 unsupported unit types. Jobs enqueued for unsupported unit
4737 types will now fail with an "unsupported" error code. More
4738 specifically .swap, .automount and .device units are not
4739 supported in containers, .busname units are not supported on
4740 non-kdbus systems. .swap and .automount are also not
4741 supported if their respective kernel compile time options
4744 * machinectl gained support for two new "copy-from" and
4745 "copy-to" commands for copying files from a running
4746 container to the host or vice versa.
4748 * machinectl gained support for a new "bind" command to bind
4749 mount host directories into local containers. This is
4750 currently only supported for nspawn containers.
4752 * networkd gained support for configuring bridge forwarding
4753 database entries (fdb) from .network files.
4755 * A new tiny daemon "systemd-importd" has been added that can
4756 download container images in tar, raw, qcow2 or dkr formats,
4757 and make them available locally in /var/lib/machines, so
4758 that they can run as nspawn containers. The daemon can GPG
4759 verify the downloads (not supported for dkr, since it has no
4760 provisions for verifying downloads). It will transparently
4761 decompress bz2, xz, gzip compressed downloads if necessary,
4762 and restore sparse files on disk. The daemon uses privilege
4763 separation to ensure the actual download logic runs with
4764 fewer privileges than the daemon itself. machinectl has
4765 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
4766 make the functionality of importd available to the
4767 user. With this in place the Fedora and Ubuntu "Cloud"
4768 images can be downloaded and booted as containers unmodified
4769 (the Fedora images lack the appropriate GPG signature files
4770 currently, so they cannot be verified, but this will change
4771 soon, hopefully). Note that downloading images is currently
4772 only fully supported on btrfs.
4774 * machinectl is now able to list container images found in
4775 /var/lib/machines, along with some metadata about sizes of
4776 disk and similar. If the directory is located on btrfs and
4777 quota is enabled, this includes quota display. A new command
4778 "image-status" has been added that shows additional
4779 information about images.
4781 * machinectl is now able to clone container images
4782 efficiently, if the underlying file system (btrfs) supports
4783 it, with the new "machinectl clone" command. It also
4784 gained commands for renaming and removing images, as well as
4785 marking them read-only or read-write (supported also on
4786 legacy file systems).
4788 * networkd gained support for collecting LLDP network
4789 announcements, from hardware that supports this. This is
4790 shown in networkctl output.
4792 * systemd-run gained support for a new -t (--pty) switch for
4793 invoking a binary on a pty whose input and output is
4794 connected to the invoking terminal. This allows executing
4795 processes as system services while interactively
4796 communicating with them via the terminal. Most interestingly
4797 this is supported across container boundaries. Invoking
4798 "systemd-run -t /bin/bash" is an alternative to running a
4799 full login session, the difference being that the former
4800 will not register a session, nor go through the PAM session
4803 * tmpfiles gained support for a new "v" line type for creating
4804 btrfs subvolumes. If the underlying file system is a legacy
4805 file system, this automatically degrades to creating a
4806 normal directory. Among others /var/lib/machines is now
4807 created like this at boot, should it be missing.
4809 * The directory /var/lib/containers/ has been deprecated and
4810 been replaced by /var/lib/machines. The term "machines" has
4811 been used in the systemd context as generic term for both
4812 VMs and containers, and hence appears more appropriate for
4813 this, as the directory can also contain raw images bootable
4816 * systemd-nspawn when invoked with -M but without --directory=
4817 or --image= is now capable of searching for the container
4818 root directory, subvolume or disk image automatically, in
4819 /var/lib/machines. systemd-nspawn@.service has been updated
4820 to make use of this, thus allowing it to be used for raw
4823 * A new machines.target unit has been introduced that is
4824 supposed to group all containers/VMs invoked as services on
4825 the system. systemd-nspawn@.service has been updated to
4826 integrate with that.
4828 * machinectl gained a new "start" command, for invoking a
4829 container as a service. "machinectl start foo" is mostly
4830 equivalent to "systemctl start systemd-nspawn@foo.service",
4831 but handles escaping in a nicer way.
4833 * systemd-nspawn will now mount most of the cgroupfs tree
4834 read-only into each container, with the exception of the
4835 container's own subtree in the name=systemd hierarchy.
4837 * journald now sets the special FS_NOCOW file flag for its
4838 journal files. This should improve performance on btrfs, by
4839 avoiding heavy fragmentation when journald's write-pattern
4840 is used on COW file systems. It degrades btrfs' data
4841 integrity guarantees for the files to the same levels as for
4842 ext3/ext4 however. This should be OK though as journald does
4843 its own data integrity checks and all its objects are
4844 checksummed on disk. Also, journald should handle btrfs disk
4845 full events a lot more gracefully now, by processing SIGBUS
4846 errors, and not relying on fallocate() anymore.
4848 * When journald detects that journal files it is writing to
4849 have been deleted it will immediately start new journal
4852 * systemd now provides a way to store file descriptors
4853 per-service in PID 1. This is useful for daemons to ensure
4854 that fds they require are not lost during a daemon
4855 restart. The fds are passed to the daemon on the next
4856 invocation in the same way socket activation fds are
4857 passed. This is now used by journald to ensure that the
4858 various sockets connected to all the system's stdout/stderr
4859 are not lost when journald is restarted. File descriptors
4860 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
4861 an extension to sd_notify(). Note that a limit is enforced
4862 on the number of fds a service can store in PID 1, and it
4863 defaults to 0, so that no fds may be stored, unless this is
4864 explicitly turned on.
4866 * The default TERM variable to use for units connected to a
4867 terminal, when no other value is explicitly is set is now
4868 vt220 rather than vt102. This should be fairly safe still,
4869 but allows PgUp/PgDn work.
4871 * The /etc/crypttab option header= as known from Debian is now
4874 * "loginctl user-status" and "loginctl session-status" will
4875 now show the last 10 lines of log messages of the
4876 user/session following the status output. Similar,
4877 "machinectl status" will show the last 10 log lines
4878 associated with a virtual machine or container
4879 service. (Note that this is usually not the log messages
4880 done in the VM/container itself, but simply what the
4881 container manager logs. For nspawn this includes all console
4884 * "loginctl session-status" without further argument will now
4885 show the status of the session of the caller. Similar,
4886 "lock-session", "unlock-session", "activate",
4887 "enable-linger", "disable-linger" may now be called without
4888 session/user parameter in which case they apply to the
4889 caller's session/user.
4891 * An X11 session scriptlet is now shipped that uploads
4892 $DISPLAY and $XAUTHORITY into the environment of the systemd
4893 --user daemon if a session begins. This should improve
4894 compatibility with X11 enabled applications run as systemd
4897 * Generators are now subject to masking via /etc and /run, the
4898 same way as unit files.
4900 * networkd .network files gained support for configuring
4901 per-link IPv4/IPv6 packet forwarding as well as IPv4
4902 masquerading. This is by default turned on for veth links to
4903 containers, as registered by systemd-nspawn. This means that
4904 nspawn containers run with --network-veth will now get
4905 automatic routed access to the host's networks without any
4906 further configuration or setup, as long as networkd runs on
4909 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4910 or UDP posts of a container on the host. With this in place
4911 it is possible to run containers with private veth links
4912 (--network-veth), and have their functionality exposed on
4913 the host as if their services were running directly on the
4916 * systemd-nspawn's --network-veth switch now gained a short
4917 version "-n", since with the changes above it is now truly
4918 useful out-of-the-box. The systemd-nspawn@.service has been
4919 updated to make use of it too by default.
4921 * systemd-nspawn will now maintain a per-image R/W lock, to
4922 ensure that the same image is not started more than once
4923 writable. (It's OK to run an image multiple times
4924 simultaneously in read-only mode.)
4926 * systemd-nspawn's --image= option is now capable of
4927 dissecting and booting MBR and GPT disk images that contain
4928 only a single active Linux partition. Previously it
4929 supported only GPT disk images with proper GPT type
4930 IDs. This allows running cloud images from major
4931 distributions directly with systemd-nspawn, without
4934 * In addition to collecting mouse dpi data in the udev
4935 hardware database, there's now support for collecting angle
4936 information for mouse scroll wheels. The database is
4937 supposed to guarantee similar scrolling behavior on mice
4938 that it knows about. There's also support for collecting
4939 information about Touchpad types.
4941 * udev's input_id built-in will now also collect touch screen
4942 dimension data and attach it to probed devices.
4944 * /etc/os-release gained support for a Distribution Privacy
4947 * networkd gained support for creating "ipvlan", "gretap",
4948 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4950 * systemd-tmpfiles gained support for "a" lines for setting
4953 * systemd-nspawn will now mount /tmp in the container to
4954 tmpfs, automatically.
4956 * systemd now exposes the memory.usage_in_bytes cgroup
4957 attribute and shows it for each service in the "systemctl
4958 status" output, if available.
4960 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4961 immediate reboot is triggered. This useful if shutdown is
4962 hung and is unable to complete, to expedite the
4963 operation. Note that this kind of reboot will still unmount
4964 all file systems, and hence should not result in fsck being
4967 * A .device unit for an optical block device will now be
4968 considered active only when a medium is in the drive. Also,
4969 mount units are now bound to their backing devices thus
4970 triggering automatic unmounting when devices become
4971 unavailable. With this in place systemd will now
4972 automatically unmount left-over mounts when a CD-ROM is
4973 ejected or an USB stick is yanked from the system.
4975 * networkd-wait-online now has support for waiting for
4976 specific interfaces only (with globbing), and for giving up
4977 after a configurable timeout.
4979 * networkd now exits when idle. It will be automatically
4980 restarted as soon as interfaces show up, are removed or
4981 change state. networkd will stay around as long as there is
4982 at least one DHCP state machine or similar around, that keep
4985 * networkd may now configure IPv6 link-local addressing in
4986 addition to IPv4 link-local addressing.
4988 * The IPv6 "token" for use in SLAAC may now be configured for
4989 each .network interface in networkd.
4991 * Routes configured with networkd may now be assigned a scope
4994 * networkd's [Match] sections now support globbing and lists
4995 of multiple space-separated matches per item.
4997 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
4998 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4999 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
5000 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
5001 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
5002 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
5003 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
5004 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
5005 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
5006 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
5007 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
5008 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
5009 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
5010 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
5011 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
5012 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
5013 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
5014 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
5015 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
5016 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
5017 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
5018 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
5019 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
5020 Hoffmann, Zbigniew Jędrzejewski-Szmek
5022 — Berlin, 2015-02-16
5026 * When querying unit file enablement status (for example via
5027 "systemctl is-enabled"), a new state "indirect" is now known
5028 which indicates that a unit might not be enabled itself, but
5029 another unit listed in its Also= setting might be.
5031 * Similar to the various existing ConditionXYZ= settings for
5032 units, there are now matching AssertXYZ= settings. While
5033 failing conditions cause a unit to be skipped, but its job
5034 to succeed, failing assertions declared like this will cause
5035 a unit start operation and its job to fail.
5037 * hostnamed now knows a new chassis type "embedded".
5039 * systemctl gained a new "edit" command. When used on a unit
5040 file, this allows extending unit files with .d/ drop-in
5041 configuration snippets or editing the full file (after
5042 copying it from /usr/lib to /etc). This will invoke the
5043 user's editor (as configured with $EDITOR), and reload the
5044 modified configuration after editing.
5046 * "systemctl status" now shows the suggested enablement state
5047 for a unit, as declared in the (usually vendor-supplied)
5048 system preset files.
5050 * nss-myhostname will now resolve the single-label host name
5051 "gateway" to the locally configured default IP routing
5052 gateways, ordered by their metrics. This assigns a stable
5053 name to the used gateways, regardless which ones are
5054 currently configured. Note that the name will only be
5055 resolved after all other name sources (if nss-myhostname is
5056 configured properly) and should hence not negatively impact
5057 systems that use the single-label host name "gateway" in
5060 * systemd-inhibit now allows filtering by mode when listing
5063 * Scope and service units gained a new "Delegate" boolean
5064 property, which, when set, allows processes running inside the
5065 unit to further partition resources. This is primarily
5066 useful for systemd user instances as well as container
5069 * journald will now pick up audit messages directly from
5070 the kernel, and log them like any other log message. The
5071 audit fields are split up and fully indexed. This means that
5072 journalctl in many ways is now a (nicer!) alternative to
5073 ausearch, the traditional audit client. Note that this
5074 implements only a minimal audit client. If you want the
5075 special audit modes like reboot-on-log-overflow, please use
5076 the traditional auditd instead, which can be used in
5077 parallel to journald.
5079 * The ConditionSecurity= unit file option now understands the
5080 special string "audit" to check whether auditing is
5083 * journalctl gained two new commands --vacuum-size= and
5084 --vacuum-time= to delete old journal files until the
5085 remaining ones take up no more than the specified size on disk,
5086 or are not older than the specified time.
5088 * A new, native PPPoE library has been added to sd-network,
5089 systemd's library of light-weight networking protocols. This
5090 library will be used in a future version of networkd to
5091 enable PPPoE communication without an external pppd daemon.
5093 * The busctl tool now understands a new "capture" verb that
5094 works similar to "monitor", but writes a packet capture
5095 trace to STDOUT that can be redirected to a file which is
5096 compatible with libcap's capture file format. This can then
5097 be loaded in Wireshark and similar tools to inspect bus
5100 * The busctl tool now understands a new "tree" verb that shows
5101 the object trees of a specific service on the bus, or of all
5104 * The busctl tool now understands a new "introspect" verb that
5105 shows all interfaces and members of objects on the bus,
5106 including their signature and values. This is particularly
5107 useful to get more information about bus objects shown by
5108 the new "busctl tree" command.
5110 * The busctl tool now understands new verbs "call",
5111 "set-property" and "get-property" for invoking bus method
5112 calls, setting and getting bus object properties in a
5115 * busctl gained a new --augment-creds= argument that controls
5116 whether the tool shall augment credential information it
5117 gets from the bus with data from /proc, in a possibly
5120 * nspawn's --link-journal= switch gained two new values
5121 "try-guest" and "try-host" that work like "guest" and
5122 "host", but do not fail if the host has no persistent
5123 journalling enabled. -j is now equivalent to
5124 --link-journal=try-guest.
5126 * macvlan network devices created by nspawn will now have
5127 stable MAC addresses.
5129 * A new SmackProcessLabel= unit setting has been added, which
5130 controls the SMACK security label processes forked off by
5131 the respective unit shall use.
5133 * If compiled with --enable-xkbcommon, systemd-localed will
5134 verify x11 keymap settings by compiling the given keymap. It
5135 will spew out warnings if the compilation fails. This
5136 requires libxkbcommon to be installed.
5138 * When a coredump is collected, a larger number of metadata
5139 fields is now collected and included in the journal records
5140 created for it. More specifically, control group membership,
5141 environment variables, memory maps, working directory,
5142 chroot directory, /proc/$PID/status, and a list of open file
5143 descriptors is now stored in the log entry.
5145 * The udev hwdb now contains DPI information for mice. For
5148 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
5150 * All systemd programs that read standalone configuration
5151 files in /etc now also support a corresponding series of
5152 .conf.d configuration directories in /etc/, /run/,
5153 /usr/local/lib/, /usr/lib/, and (if configured with
5154 --enable-split-usr) /lib/. In particular, the following
5155 configuration files now have corresponding configuration
5156 directories: system.conf user.conf, logind.conf,
5157 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
5158 resolved.conf, timesyncd.conf, journal-remote.conf, and
5159 journal-upload.conf. Note that distributions should use the
5160 configuration directories in /usr/lib/; the directories in
5161 /etc/ are reserved for the system administrator.
5163 * systemd-rfkill will no longer take the rfkill device name
5164 into account when storing rfkill state on disk, as the name
5165 might be dynamically assigned and not stable. Instead, the
5166 ID_PATH udev variable combined with the rfkill type (wlan,
5167 bluetooth, ...) is used.
5169 * A new service systemd-machine-id-commit.service has been
5170 added. When used on systems where /etc is read-only during
5171 boot, and /etc/machine-id is not initialized (but an empty
5172 file), this service will copy the temporary machine ID
5173 created as replacement into /etc after the system is fully
5174 booted up. This is useful for systems that are freshly
5175 installed with a non-initialized machine ID, but should get
5176 a fixed machine ID for subsequent boots.
5178 * networkd's .netdev files now provide a large set of
5179 configuration parameters for VXLAN devices. Similarly, the
5180 bridge port cost parameter is now configurable in .network
5181 files. There's also new support for configuring IP source
5182 routing. networkd .link files gained support for a new
5183 OriginalName= match that is useful to match against the
5184 original interface name the kernel assigned. .network files
5185 may include MTU= and MACAddress= fields for altering the MTU
5186 and MAC address while being connected to a specific network
5189 * The LUKS logic gained supported for configuring
5190 UUID-specific key files. There's also new support for naming
5191 LUKS device from the kernel command line, using the new
5192 luks.name= argument.
5194 * Timer units may now be transiently created via the bus API
5195 (this was previously already available for scope and service
5196 units). In addition it is now possible to create multiple
5197 transient units at the same time with a single bus call. The
5198 "systemd-run" tool has been updated to make use of this for
5199 running commands on a specified time, in at(1)-style.
5201 * tmpfiles gained support for "t" lines, for assigning
5202 extended attributes to files. Among other uses this may be
5203 used to assign SMACK labels to files.
5205 Contributions from: Alin Rauta, Alison Chaiken, Andrej
5206 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
5207 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
5208 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
5209 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
5210 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
5211 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
5212 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
5213 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
5214 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
5215 Michael Biebl, Michael Chapman, Michael Marineau, Michal
5216 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
5217 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
5218 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
5219 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
5220 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5221 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
5222 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
5224 — Berlin, 2014-12-10
5228 * journalctl gained the new options -t/--identifier= to match
5229 on the syslog identifier (aka "tag"), as well as --utc to
5230 show log timestamps in the UTC timezone. journalctl now also
5231 accepts -n/--lines=all to disable line capping in a pager.
5233 * journalctl gained a new switch, --flush, that synchronously
5234 flushes logs from /run/log/journal to /var/log/journal if
5235 persistent storage is enabled. systemd-journal-flush.service
5236 now waits until the operation is complete.
5238 * Services can notify the manager before they start a reload
5239 (by sending RELOADING=1) or shutdown (by sending
5240 STOPPING=1). This allows the manager to track and show the
5241 internal state of daemons and closes a race condition when
5242 the process is still running but has closed its D-Bus
5245 * Services with Type=oneshot do not have to have any ExecStart
5248 * User units are now loaded also from
5249 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
5250 /run/systemd/user directory that was already previously
5251 supported, but is under the control of the user.
5253 * Job timeouts (i.e. timeouts on the time a job that is
5254 queued stays in the run queue) can now optionally result in
5255 immediate reboot or power-off actions (JobTimeoutAction= and
5256 JobTimeoutRebootArgument=). This is useful on ".target"
5257 units, to limit the maximum time a target remains
5258 undispatched in the run queue, and to trigger an emergency
5259 operation in such a case. This is now used by default to
5260 turn off the system if boot-up (as defined by everything in
5261 basic.target) hangs and does not complete for at least
5262 15min. Also, if power-off or reboot hang for at least 30min
5263 an immediate power-off/reboot operation is triggered. This
5264 functionality is particularly useful to increase reliability
5265 on embedded devices, but also on laptops which might
5266 accidentally get powered on when carried in a backpack and
5267 whose boot stays stuck in a hard disk encryption passphrase
5270 * systemd-logind can be configured to also handle lid switch
5271 events even when the machine is docked or multiple displays
5272 are attached (HandleLidSwitchDocked= option).
5274 * A helper binary and a service have been added which can be
5275 used to resume from hibernation in the initramfs. A
5276 generator will parse the resume= option on the kernel
5277 command line to trigger resume.
5279 * A user console daemon systemd-consoled has been
5280 added. Currently, it is a preview, and will so far open a
5281 single terminal on each session of the user marked as
5282 Desktop=systemd-console.
5284 * Route metrics can be specified for DHCP routes added by
5287 * The SELinux context of socket-activated services can be set
5288 from the information provided by the networking stack
5289 (SELinuxContextFromNet= option).
5291 * Userspace firmware loading support has been removed and
5292 the minimum supported kernel version is thus bumped to 3.7.
5294 * Timeout for udev workers has been increased from 1 to 3
5295 minutes, but a warning will be printed after 1 minute to
5296 help diagnose kernel modules that take a long time to load.
5298 * Udev rules can now remove tags on devices with TAG-="foobar".
5300 * systemd's readahead implementation has been removed. In many
5301 circumstances it didn't give expected benefits even for
5302 rotational disk drives and was becoming less relevant in the
5303 age of SSDs. As none of the developers has been using
5304 rotating media anymore, and nobody stepped up to actively
5305 maintain this component of systemd it has now been removed.
5307 * Swap units can use Options= to specify discard options.
5308 Discard options specified for swaps in /etc/fstab are now
5311 * Docker containers are now detected as a separate type of
5314 * The Password Agent protocol gained support for queries where
5315 the user input is shown, useful e.g. for user names.
5316 systemd-ask-password gained a new --echo option to turn that
5319 * The default sysctl.d/ snippets will now set:
5321 net.core.default_qdisc = fq_codel
5323 This selects Fair Queuing Controlled Delay as the default
5324 queuing discipline for network interfaces. fq_codel helps
5325 fight the network bufferbloat problem. It is believed to be
5326 a good default with no tuning required for most workloads.
5327 Downstream distributions may override this choice. On 10Gbit
5328 servers that do not do forwarding, "fq" may perform better.
5329 Systems without a good clocksource should use "pfifo_fast".
5331 * If kdbus is enabled during build a new option BusPolicy= is
5332 available for service units, that allows locking all service
5333 processes into a stricter bus policy, in order to limit
5334 access to various bus services, or even hide most of them
5335 from the service's view entirely.
5337 * networkctl will now show the .network and .link file
5338 networkd has applied to a specific interface.
5340 * sd-login gained a new API call sd_session_get_desktop() to
5341 query which desktop environment has been selected for a
5344 * UNIX utmp support is now compile-time optional to support
5345 legacy-free systems.
5347 * systemctl gained two new commands "add-wants" and
5348 "add-requires" for pulling in units from specific targets
5351 * If the word "rescue" is specified on the kernel command line
5352 the system will now boot into rescue mode (aka
5353 rescue.target), which was previously available only by
5354 specifying "1" or "systemd.unit=rescue.target" on the kernel
5355 command line. This new kernel command line option nicely
5356 mirrors the already existing "emergency" kernel command line
5359 * New kernel command line options mount.usr=, mount.usrflags=,
5360 mount.usrfstype= have been added that match root=, rootflags=,
5361 rootfstype= but allow mounting a specific file system to
5364 * The $NOTIFY_SOCKET is now also passed to control processes of
5365 services, not only the main process.
5367 * This version reenables support for fsck's -l switch. This
5368 means at least version v2.25 of util-linux is required for
5369 operation, otherwise dead-locks on device nodes may
5370 occur. Again: you need to update util-linux to at least
5371 v2.25 when updating systemd to v217.
5373 * The "multi-seat-x" tool has been removed from systemd, as
5374 its functionality has been integrated into X servers 1.16,
5375 and the tool is hence redundant. It is recommended to update
5376 display managers invoking this tool to simply invoke X
5377 directly from now on, again.
5379 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
5380 message flag has been added for all of systemd's polkit
5381 authenticated method calls has been added. In particular this
5382 now allows optional interactive authorization via polkit for
5383 many of PID1's privileged operations such as unit file
5384 enabling and disabling.
5386 * "udevadm hwdb --update" learnt a new switch "--usr" for
5387 placing the rebuilt hardware database in /usr instead of
5388 /etc. When used only hardware database entries stored in
5389 /usr will be used, and any user database entries in /etc are
5390 ignored. This functionality is useful for vendors to ship a
5391 pre-built database on systems where local configuration is
5392 unnecessary or unlikely.
5394 * Calendar time specifications in .timer units now also
5395 understand the strings "semi-annually", "quarterly" and
5396 "minutely" as shortcuts (in addition to the preexisting
5397 "anually", "hourly", ...).
5399 * systemd-tmpfiles will now correctly create files in /dev
5400 at boot which are marked for creation only at boot. It is
5401 recommended to always create static device nodes with 'c!'
5402 and 'b!', so that they are created only at boot and not
5403 overwritten at runtime.
5405 * When the watchdog logic is used for a service (WatchdogSec=)
5406 and the watchdog timeout is hit the service will now be
5407 terminated with SIGABRT (instead of just SIGTERM), in order
5408 to make sure a proper coredump and backtrace is
5409 generated. This ensures that hanging services will result in
5410 similar coredump/backtrace behaviour as services that hit a
5413 Contributions from: Andreas Henriksson, Andrei Borzenkov,
5414 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
5415 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
5416 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
5417 Herrmann, David Sommerseth, David Strauss, Emil Renner
5418 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
5419 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
5420 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
5421 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
5422 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
5423 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
5424 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
5425 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
5426 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
5427 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
5428 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
5429 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
5430 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
5431 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
5432 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
5433 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
5436 — Berlin, 2014-10-28
5440 * timedated no longer reads NTP implementation unit names from
5441 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
5442 implementations should add a
5444 Conflicts=systemd-timesyncd.service
5446 to their unit files to take over and replace systemd's NTP
5447 default functionality.
5449 * systemd-sysusers gained a new line type "r" for configuring
5450 which UID/GID ranges to allocate system users/groups
5451 from. Lines of type "u" may now add an additional column
5452 that specifies the home directory for the system user to be
5453 created. Also, systemd-sysusers may now optionally read user
5454 information from STDIN instead of a file. This is useful for
5455 invoking it from RPM preinst scriptlets that need to create
5456 users before the first RPM file is installed since these
5457 files might need to be owned by them. A new
5458 %sysusers_create_inline RPM macro has been introduced to do
5459 just that. systemd-sysusers now updates the shadow files as
5460 well as the user/group databases, which should enhance
5461 compatibility with certain tools like grpck.
5463 * A number of bus APIs of PID 1 now optionally consult polkit to
5464 permit access for otherwise unprivileged clients under certain
5465 conditions. Note that this currently doesn't support
5466 interactive authentication yet, but this is expected to be
5467 added eventually, too.
5469 * /etc/machine-info now has new fields for configuring the
5470 deployment environment of the machine, as well as the
5471 location of the machine. hostnamectl has been updated with
5472 new command to update these fields.
5474 * systemd-timesyncd has been updated to automatically acquire
5475 NTP server information from systemd-networkd, which might
5476 have been discovered via DHCP.
5478 * systemd-resolved now includes a caching DNS stub resolver
5479 and a complete LLMNR name resolution implementation. A new
5480 NSS module "nss-resolve" has been added which can be used
5481 instead of glibc's own "nss-dns" to resolve hostnames via
5482 systemd-resolved. Hostnames, addresses and arbitrary RRs may
5483 be resolved via systemd-resolved D-Bus APIs. In contrast to
5484 the glibc internal resolver systemd-resolved is aware of
5485 multi-homed system, and keeps DNS server and caches separate
5486 and per-interface. Queries are sent simultaneously on all
5487 interfaces that have DNS servers configured, in order to
5488 properly handle VPNs and local LANs which might resolve
5489 separate sets of domain names. systemd-resolved may acquire
5490 DNS server information from systemd-networkd automatically,
5491 which in turn might have discovered them via DHCP. A tool
5492 "systemd-resolve-host" has been added that may be used to
5493 query the DNS logic in resolved. systemd-resolved implements
5494 IDNA and automatically uses IDNA or UTF-8 encoding depending
5495 on whether classic DNS or LLMNR is used as transport. In the
5496 next releases we intend to add a DNSSEC and mDNS/DNS-SD
5497 implementation to systemd-resolved.
5499 * A new NSS module nss-mymachines has been added, that
5500 automatically resolves the names of all local registered
5501 containers to their respective IP addresses.
5503 * A new client tool "networkctl" for systemd-networkd has been
5504 added. It currently is entirely passive and will query
5505 networking configuration from udev, rtnetlink and networkd,
5506 and present it to the user in a very friendly
5507 way. Eventually, we hope to extend it to become a full
5508 control utility for networkd.
5510 * .socket units gained a new DeferAcceptSec= setting that
5511 controls the kernels' TCP_DEFER_ACCEPT sockopt for
5512 TCP. Similarly, support for controlling TCP keep-alive
5513 settings has been added (KeepAliveTimeSec=,
5514 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
5515 turning off Nagle's algorithm on TCP has been added
5518 * logind learned a new session type "web", for use in projects
5519 like Cockpit which register web clients as PAM sessions.
5521 * timer units with at least one OnCalendar= setting will now
5522 be started only after time-sync.target has been
5523 reached. This way they will not elapse before the system
5524 clock has been corrected by a local NTP client or
5525 similar. This is particular useful on RTC-less embedded
5526 machines, that come up with an invalid system clock.
5528 * systemd-nspawn's --network-veth= switch should now result in
5529 stable MAC addresses for both the outer and the inner side
5532 * systemd-nspawn gained a new --volatile= switch for running
5533 container instances with /etc or /var unpopulated.
5535 * The kdbus client code has been updated to use the new Linux
5536 3.17 memfd subsystem instead of the old kdbus-specific one.
5538 * systemd-networkd's DHCP client and server now support
5539 FORCERENEW. There are also new configuration options to
5540 configure the vendor client identifier and broadcast mode
5543 * systemd will no longer inform the kernel about the current
5544 timezone, as this is necessarily incorrect and racy as the
5545 kernel has no understanding of DST and similar
5546 concepts. This hence means FAT timestamps will be always
5547 considered UTC, similar to what Android is already
5548 doing. Also, when the RTC is configured to the local time
5549 (rather than UTC) systemd will never synchronize back to it,
5550 as this might confuse Windows at a later boot.
5552 * systemd-analyze gained a new command "verify" for offline
5553 validation of unit files.
5555 * systemd-networkd gained support for a couple of additional
5556 settings for bonding networking setups. Also, the metric for
5557 statically configured routes may now be configured. For
5558 network interfaces where this is appropriate the peer IP
5559 address may now be configured.
5561 * systemd-networkd's DHCP client will no longer request
5562 broadcasting by default, as this tripped up some networks.
5563 For hardware where broadcast is required the feature should
5564 be switched back on using RequestBroadcast=yes.
5566 * systemd-networkd will now set up IPv4LL addresses (when
5567 enabled) even if DHCP is configured successfully.
5569 * udev will now default to respect network device names given
5570 by the kernel when the kernel indicates that these are
5571 predictable. This behavior can be tweaked by changing
5572 NamePolicy= in the relevant .link file.
5574 * A new library systemd-terminal has been added that
5575 implements full TTY stream parsing and rendering. This
5576 library is supposed to be used later on for implementing a
5577 full userspace VT subsystem, replacing the current kernel
5580 * A new tool systemd-journal-upload has been added to push
5581 journal data to a remote system running
5582 systemd-journal-remote.
5584 * journald will no longer forward all local data to another
5585 running syslog daemon. This change has been made because
5586 rsyslog (which appears to be the most commonly used syslog
5587 implementation these days) no longer makes use of this, and
5588 instead pulls the data out of the journal on its own. Since
5589 forwarding the messages to a non-existent syslog server is
5590 more expensive than we assumed we have now turned this
5591 off. If you run a syslog server that is not a recent rsyslog
5592 version, you have to turn this option on again
5593 (ForwardToSyslog= in journald.conf).
5595 * journald now optionally supports the LZ4 compressor for
5596 larger journal fields. This compressor should perform much
5597 better than XZ which was the previous default.
5599 * machinectl now shows the IP addresses of local containers,
5600 if it knows them, plus the interface name of the container.
5602 * A new tool "systemd-escape" has been added that makes it
5603 easy to escape strings to build unit names and similar.
5605 * sd_notify() messages may now include a new ERRNO= field
5606 which is parsed and collected by systemd and shown among the
5607 "systemctl status" output for a service.
5609 * A new component "systemd-firstboot" has been added that
5610 queries the most basic systemd information (timezone,
5611 hostname, root password) interactively on first
5612 boot. Alternatively it may also be used to provision these
5613 things offline on OS images installed into directories.
5615 * The default sysctl.d/ snippets will now set
5617 net.ipv4.conf.default.promote_secondaries=1
5619 This has the benefit of no flushing secondary IP addresses
5620 when primary addresses are removed.
5622 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
5623 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
5624 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
5625 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
5626 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
5627 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
5628 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
5629 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
5630 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
5631 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
5632 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
5633 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
5634 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
5635 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
5636 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
5638 — Berlin, 2014-08-19
5642 * A new tool systemd-sysusers has been added. This tool
5643 creates system users and groups in /etc/passwd and
5644 /etc/group, based on static declarative system user/group
5645 definitions in /usr/lib/sysusers.d/. This is useful to
5646 enable factory resets and volatile systems that boot up with
5647 an empty /etc directory, and thus need system users and
5648 groups created during early boot. systemd now also ships
5649 with two default sysusers.d/ files for the most basic
5650 users and groups systemd and the core operating system
5653 * A new tmpfiles snippet has been added that rebuilds the
5654 essential files in /etc on boot, should they be missing.
5656 * A directive for ensuring automatic clean-up of
5657 /var/cache/man/ has been removed from the default
5658 configuration. This line should now be shipped by the man
5659 implementation. The necessary change has been made to the
5660 man-db implementation. Note that you need to update your man
5661 implementation to one that ships this line, otherwise no
5662 automatic clean-up of /var/cache/man will take place.
5664 * A new condition ConditionNeedsUpdate= has been added that
5665 may conditionalize services to only run when /etc or /var
5666 are "older" than the vendor operating system resources in
5667 /usr. This is useful for reconstructing or updating /etc
5668 after an offline update of /usr or a factory reset, on the
5669 next reboot. Services that want to run once after such an
5670 update or reset should use this condition and order
5671 themselves before the new systemd-update-done.service, which
5672 will mark the two directories as fully updated. A number of
5673 service files have been added making use of this, to rebuild
5674 the udev hardware database, the journald message catalog and
5675 dynamic loader cache (ldconfig). The systemd-sysusers tool
5676 described above also makes use of this now. With this in
5677 place it is now possible to start up a minimal operating
5678 system with /etc empty cleanly. For more information on the
5679 concepts involved see this recent blog story:
5681 http://0pointer.de/blog/projects/stateless.html
5683 * A new system group "input" has been introduced, and all
5684 input device nodes get this group assigned. This is useful
5685 for system-level software to get access to input devices. It
5686 complements what is already done for "audio" and "video".
5688 * systemd-networkd learnt minimal DHCPv4 server support in
5689 addition to the existing DHCPv4 client support. It also
5690 learnt DHCPv6 client and IPv6 Router Solicitation client
5691 support. The DHCPv4 client gained support for static routes
5692 passed in from the server. Note that the [DHCPv4] section
5693 known in older systemd-networkd versions has been renamed to
5694 [DHCP] and is now also used by the DHCPv6 client. Existing
5695 .network files using settings of this section should be
5696 updated, though compatibility is maintained. Optionally, the
5697 client hostname may now be sent to the DHCP server.
5699 * networkd gained support for vxlan virtual networks as well
5700 as tun/tap and dummy devices.
5702 * networkd gained support for automatic allocation of address
5703 ranges for interfaces from a system-wide pool of
5704 addresses. This is useful for dynamically managing a large
5705 number of interfaces with a single network configuration
5706 file. In particular this is useful to easily assign
5707 appropriate IP addresses to the veth links of a large number
5708 of nspawn instances.
5710 * RPM macros for processing sysusers, sysctl and binfmt
5711 drop-in snippets at package installation time have been
5714 * The /etc/os-release file should now be placed in
5715 /usr/lib/os-release. The old location is automatically
5716 created as symlink. /usr/lib is the more appropriate
5717 location of this file, since it shall actually describe the
5718 vendor operating system shipped in /usr, and not the
5719 configuration stored in /etc.
5721 * .mount units gained a new boolean SloppyOptions= setting
5722 that maps to mount(8)'s -s option which enables permissive
5723 parsing of unknown mount options.
5725 * tmpfiles learnt a new "L+" directive which creates a symlink
5726 but (unlike "L") deletes a pre-existing file first, should
5727 it already exist and not already be the correct
5728 symlink. Similarly, "b+", "c+" and "p+" directives have been
5729 added as well, which create block and character devices, as
5730 well as fifos in the filesystem, possibly removing any
5731 pre-existing files of different types.
5733 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
5734 'argument' field (which so far specified the source to
5735 symlink/copy the files from) is now optional. If omitted the
5736 same file os copied from /usr/share/factory/ suffixed by the
5737 full destination path. This is useful for populating /etc
5738 with essential files, by copying them from vendor defaults
5739 shipped in /usr/share/factory/etc.
5741 * A new command "systemctl preset-all" has been added that
5742 applies the service preset settings to all installed unit
5743 files. A new switch --preset-mode= has been added that
5744 controls whether only enable or only disable operations
5747 * A new command "systemctl is-system-running" has been added
5748 that allows checking the overall state of the system, for
5749 example whether it is fully up and running.
5751 * When the system boots up with an empty /etc, the equivalent
5752 to "systemctl preset-all" is executed during early boot, to
5753 make sure all default services are enabled after a factory
5756 * systemd now contains a minimal preset file that enables the
5757 most basic services systemd ships by default.
5759 * Unit files' [Install] section gained a new DefaultInstance=
5760 field for defining the default instance to create if a
5761 template unit is enabled with no instance specified.
5763 * A new passive target cryptsetup-pre.target has been added
5764 that may be used by services that need to make they run and
5765 finish before the first LUKS cryptographic device is set up.
5767 * The /dev/loop-control and /dev/btrfs-control device nodes
5768 are now owned by the "disk" group by default, opening up
5769 access to this group.
5771 * systemd-coredump will now automatically generate a
5772 stack trace of all core dumps taking place on the system,
5773 based on elfutils' libdw library. This stack trace is logged
5776 * systemd-coredump may now optionally store coredumps directly
5777 on disk (in /var/lib/systemd/coredump, possibly compressed),
5778 instead of storing them unconditionally in the journal. This
5779 mode is the new default. A new configuration file
5780 /etc/systemd/coredump.conf has been added to configure this
5781 and other parameters of systemd-coredump.
5783 * coredumpctl gained a new "info" verb to show details about a
5784 specific coredump. A new switch "-1" has also been added
5785 that makes sure to only show information about the most
5786 recent entry instead of all entries. Also, as the tool is
5787 generally useful now the "systemd-" prefix of the binary
5788 name has been removed. Distributions that want to maintain
5789 compatibility with the old name should add a symlink from
5790 the old name to the new name.
5792 * journald's SplitMode= now defaults to "uid". This makes sure
5793 that unprivileged users can access their own coredumps with
5794 coredumpctl without restrictions.
5796 * New kernel command line options "systemd.wants=" (for
5797 pulling an additional unit during boot), "systemd.mask="
5798 (for masking a specific unit for the boot), and
5799 "systemd.debug-shell" (for enabling the debug shell on tty9)
5800 have been added. This is implemented in the new generator
5801 "systemd-debug-generator".
5803 * systemd-nspawn will now by default filter a couple of
5804 syscalls for containers, among them those required for
5805 kernel module loading, direct x86 IO port access, swap
5806 management, and kexec. Most importantly though
5807 open_by_handle_at() is now prohibited for containers,
5808 closing a hole similar to a recently discussed vulnerability
5809 in docker regarding access to files on file hierarchies the
5810 container should normally not have access to. Note that, for
5811 nspawn, we generally make no security claims anyway (and
5812 this is explicitly documented in the man page), so this is
5813 just a fix for one of the most obvious problems.
5815 * A new man page file-hierarchy(7) has been added that
5816 contains a minimized, modernized version of the file system
5817 layout systemd expects, similar in style to the FHS
5818 specification or hier(5). A new tool systemd-path(1) has
5819 been added to query many of these paths for the local
5822 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
5823 longer done. Since the directory now has a per-user size
5824 limit, and is cleaned on logout this appears unnecessary,
5825 in particular since this now brings the lifecycle of this
5826 directory closer in line with how IPC objects are handled.
5828 * systemd.pc now exports a number of additional directories,
5829 including $libdir (which is useful to identify the library
5830 path for the primary architecture of the system), and a
5831 couple of drop-in directories.
5833 * udev's predictable network interface names now use the dev_port
5834 sysfs attribute, introduced in linux 3.15 instead of dev_id to
5835 distinguish between ports of the same PCI function. dev_id should
5836 only be used for ports using the same HW address, hence the need
5839 * machined has been updated to export the OS version of a
5840 container (read from /etc/os-release and
5841 /usr/lib/os-release) on the bus. This is now shown in
5842 "machinectl status" for a machine.
5844 * A new service setting RestartForceExitStatus= has been
5845 added. If configured to a set of exit signals or process
5846 return values, the service will be restarted when the main
5847 daemon process exits with any of them, regardless of the
5850 * systemctl's -H switch for connecting to remote systemd
5851 machines has been extended so that it may be used to
5852 directly connect to a specific container on the
5853 host. "systemctl -H root@foobar:waldi" will now connect as
5854 user "root" to host "foobar", and then proceed directly to
5855 the container named "waldi". Note that currently you have to
5856 authenticate as user "root" for this to work, as entering
5857 containers is a privileged operation.
5859 Contributions from: Andreas Henriksson, Benjamin Steinwender,
5860 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
5861 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
5862 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
5863 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
5864 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
5865 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
5866 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
5867 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
5868 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
5869 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
5870 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
5872 — Berlin, 2014-07-03
5876 * As an experimental feature, udev now tries to lock the
5877 disk device node (flock(LOCK_SH|LOCK_NB)) while it
5878 executes events for the disk or any of its partitions.
5879 Applications like partitioning programs can lock the
5880 disk device node (flock(LOCK_EX)) and claim temporary
5881 device ownership that way; udev will entirely skip all event
5882 handling for this disk and its partitions. If the disk
5883 was opened for writing, the close will trigger a partition
5884 table rescan in udev's "watch" facility, and if needed
5885 synthesize "change" events for the disk and all its partitions.
5886 This is now unconditionally enabled, and if it turns out to
5887 cause major problems, we might turn it on only for specific
5888 devices, or might need to disable it entirely. Device Mapper
5889 devices are excluded from this logic.
5891 * We temporarily dropped the "-l" switch for fsck invocations,
5892 since they collide with the flock() logic above. util-linux
5893 upstream has been changed already to avoid this conflict,
5894 and we will readd "-l" as soon as util-linux with this
5895 change has been released.
5897 * The dependency on libattr has been removed. Since a long
5898 time, the extended attribute calls have moved to glibc, and
5899 libattr is thus unnecessary.
5901 * Virtualization detection works without privileges now. This
5902 means the systemd-detect-virt binary no longer requires
5903 CAP_SYS_PTRACE file capabilities, and our daemons can run
5904 with fewer privileges.
5906 * systemd-networkd now runs under its own "systemd-network"
5907 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5908 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5909 loses the ability to write to files owned by root this way.
5911 * Similarly, systemd-resolved now runs under its own
5912 "systemd-resolve" user with no capabilities remaining.
5914 * Similarly, systemd-bus-proxyd now runs under its own
5915 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5917 * systemd-networkd gained support for setting up "veth"
5918 virtual Ethernet devices for container connectivity, as well
5919 as GRE and VTI tunnels.
5921 * systemd-networkd will no longer automatically attempt to
5922 manually load kernel modules necessary for certain tunnel
5923 transports. Instead, it is assumed the kernel loads them
5924 automatically when required. This only works correctly on
5925 very new kernels. On older kernels, please consider adding
5926 the kernel modules to /etc/modules-load.d/ as a work-around.
5928 * The resolv.conf file systemd-resolved generates has been
5929 moved to /run/systemd/resolve/. If you have a symlink from
5930 /etc/resolv.conf, it might be necessary to correct it.
5932 * Two new service settings, ProtectHome= and ProtectSystem=,
5933 have been added. When enabled, they will make the user data
5934 (such as /home) inaccessible or read-only and the system
5935 (such as /usr) read-only, for specific services. This allows
5936 very light-weight per-service sandboxing to avoid
5937 modifications of user data or system files from
5938 services. These two new switches have been enabled for all
5939 of systemd's long-running services, where appropriate.
5941 * Socket units gained new SocketUser= and SocketGroup=
5942 settings to set the owner user and group of AF_UNIX sockets
5943 and FIFOs in the file system.
5945 * Socket units gained a new RemoveOnStop= setting. If enabled,
5946 all FIFOS and sockets in the file system will be removed
5947 when the specific socket unit is stopped.
5949 * Socket units gained a new Symlinks= setting. It takes a list
5950 of symlinks to create to file system sockets or FIFOs
5951 created by the specific Unix sockets. This is useful to
5952 manage symlinks to socket nodes with the same lifecycle as
5955 * The /dev/log socket and /dev/initctl FIFO have been moved to
5956 /run, and have been replaced by symlinks. This allows
5957 connecting to these facilities even if PrivateDevices=yes is
5958 used for a service (which makes /dev/log itself unavailable,
5959 but /run is left). This also has the benefit of ensuring
5960 that /dev only contains device nodes, directories and
5961 symlinks, and nothing else.
5963 * sd-daemon gained two new calls sd_pid_notify() and
5964 sd_pid_notifyf(). They are similar to sd_notify() and
5965 sd_notifyf(), but allow overriding of the source PID of
5966 notification messages if permissions permit this. This is
5967 useful to send notify messages on behalf of a different
5968 process (for example, the parent process). The
5969 systemd-notify tool has been updated to make use of this
5970 when sending messages (so that notification messages now
5971 originate from the shell script invoking systemd-notify and
5972 not the systemd-notify process itself. This should minimize
5973 a race where systemd fails to associate notification
5974 messages to services when the originating process already
5977 * A new "on-abnormal" setting for Restart= has been added. If
5978 set, it will result in automatic restarts on all "abnormal"
5979 reasons for a process to exit, which includes unclean
5980 signals, core dumps, timeouts and watchdog timeouts, but
5981 does not include clean and unclean exit codes or clean
5982 signals. Restart=on-abnormal is an alternative for
5983 Restart=on-failure for services that shall be able to
5984 terminate and avoid restarts on certain errors, by
5985 indicating so with an unclean exit code. Restart=on-failure
5986 or Restart=on-abnormal is now the recommended setting for
5987 all long-running services.
5989 * If the InaccessibleDirectories= service setting points to a
5990 mount point (or if there are any submounts contained within
5991 it), it is now attempted to completely unmount it, to make
5992 the file systems truly unavailable for the respective
5995 * The ReadOnlyDirectories= service setting and
5996 systemd-nspawn's --read-only parameter are now recursively
5997 applied to all submounts, too.
5999 * Mount units may now be created transiently via the bus APIs.
6001 * The support for SysV and LSB init scripts has been removed
6002 from the systemd daemon itself. Instead, it is now
6003 implemented as a generator that creates native systemd units
6004 from these scripts when needed. This enables us to remove a
6005 substantial amount of legacy code from PID 1, following the
6006 fact that many distributions only ship a very small number
6007 of LSB/SysV init scripts nowadays.
6009 * Privileged Xen (dom0) domains are not considered
6010 virtualization anymore by the virtualization detection
6011 logic. After all, they generally have unrestricted access to
6012 the hardware and usually are used to manage the unprivileged
6015 * systemd-tmpfiles gained a new "C" line type, for copying
6016 files or entire directories.
6018 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
6019 lines. So far, they have been non-globbing versions of the
6020 latter, and have thus been redundant. In future, it is
6021 recommended to only use "z". "m" has hence been removed
6022 from the documentation, even though it stays supported.
6024 * A tmpfiles snippet to recreate the most basic structure in
6025 /var has been added. This is enough to create the /var/run →
6026 /run symlink and create a couple of structural
6027 directories. This allows systems to boot up with an empty or
6028 volatile /var. Of course, while with this change, the core OS
6029 now is capable with dealing with a volatile /var, not all
6030 user services are ready for it. However, we hope that sooner
6031 or later, many service daemons will be changed upstream so
6032 that they are able to automatically create their necessary
6033 directories in /var at boot, should they be missing. This is
6034 the first step to allow state-less systems that only require
6035 the vendor image for /usr to boot.
6037 * systemd-nspawn has gained a new --tmpfs= switch to mount an
6038 empty tmpfs instance to a specific directory. This is
6039 particularly useful for making use of the automatic
6040 reconstruction of /var (see above), by passing --tmpfs=/var.
6042 * Access modes specified in tmpfiles snippets may now be
6043 prefixed with "~", which indicates that they shall be masked
6044 by whether the existing file or directory is currently
6045 writable, readable or executable at all. Also, if specified,
6046 the sgid/suid/sticky bits will be masked for all
6049 * A new passive target unit "network-pre.target" has been
6050 added which is useful for services that shall run before any
6051 network is configured, for example firewall scripts.
6053 * The "floppy" group that previously owned the /dev/fd*
6054 devices is no longer used. The "disk" group is now used
6055 instead. Distributions should probably deprecate usage of
6058 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
6059 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
6060 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
6061 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
6062 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
6063 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
6066 — Berlin, 2014-06-11
6070 * A new "systemd-timesyncd" daemon has been added for
6071 synchronizing the system clock across the network. It
6072 implements an SNTP client. In contrast to NTP
6073 implementations such as chrony or the NTP reference server,
6074 this only implements a client side, and does not bother with
6075 the full NTP complexity, focusing only on querying time from
6076 one remote server and synchronizing the local clock to
6077 it. Unless you intend to serve NTP to networked clients or
6078 want to connect to local hardware clocks, this simple NTP
6079 client should be more than appropriate for most
6080 installations. The daemon runs with minimal privileges, and
6081 has been hooked up with networkd to only operate when
6082 network connectivity is available. The daemon saves the
6083 current clock to disk every time a new NTP sync has been
6084 acquired, and uses this to possibly correct the system clock
6085 early at bootup, in order to accommodate for systems that
6086 lack an RTC such as the Raspberry Pi and embedded devices,
6087 and to make sure that time monotonically progresses on these
6088 systems, even if it is not always correct. To make use of
6089 this daemon, a new system user and group "systemd-timesync"
6090 needs to be created on installation of systemd.
6092 * The queue "seqnum" interface of libudev has been disabled, as
6093 it was generally incompatible with device namespacing as
6094 sequence numbers of devices go "missing" if the devices are
6095 part of a different namespace.
6097 * "systemctl list-timers" and "systemctl list-sockets" gained
6098 a --recursive switch for showing units of these types also
6099 for all local containers, similar in style to the already
6100 supported --recursive switch for "systemctl list-units".
6102 * A new RebootArgument= setting has been added for service
6103 units, which may be used to specify a kernel reboot argument
6104 to use when triggering reboots with StartLimitAction=.
6106 * A new FailureAction= setting has been added for service
6107 units which may be used to specify an operation to trigger
6108 when a service fails. This works similarly to
6109 StartLimitAction=, but unlike it, controls what is done
6110 immediately rather than only after several attempts to
6111 restart the service in question.
6113 * hostnamed got updated to also expose the kernel name,
6114 release, and version on the bus. This is useful for
6115 executing commands like hostnamectl with the -H switch.
6116 systemd-analyze makes use of this to properly display
6117 details when running non-locally.
6119 * The bootchart tool can now show cgroup information in the
6120 graphs it generates.
6122 * The CFS CPU quota cgroup attribute is now exposed for
6123 services. The new CPUQuota= switch has been added for this
6124 which takes a percentage value. Setting this will have the
6125 result that a service may never get more CPU time than the
6126 specified percentage, even if the machine is otherwise idle.
6128 * systemd-networkd learned IPIP and SIT tunnel support.
6130 * LSB init scripts exposing a dependency on $network will now
6131 get a dependency on network-online.target rather than simply
6132 network.target. This should bring LSB handling closer to
6133 what it was on SysV systems.
6135 * A new fsck.repair= kernel option has been added to control
6136 how fsck shall deal with unclean file systems at boot.
6138 * The (.ini) configuration file parser will now silently
6139 ignore sections whose name begins with "X-". This may be
6140 used to maintain application-specific extension sections in unit
6143 * machined gained a new API to query the IP addresses of
6144 registered containers. "machinectl status" has been updated
6145 to show these addresses in its output.
6147 * A new call sd_uid_get_display() has been added to the
6148 sd-login APIs for querying the "primary" session of a
6149 user. The "primary" session of the user is elected from the
6150 user's sessions and generally a graphical session is
6151 preferred over a text one.
6153 * A minimal systemd-resolved daemon has been added. It
6154 currently simply acts as a companion to systemd-networkd and
6155 manages resolv.conf based on per-interface DNS
6156 configuration, possibly supplied via DHCP. In the long run
6157 we hope to extend this into a local DNSSEC enabled DNS and
6160 * The systemd-networkd-wait-online tool is now enabled by
6161 default. It will delay network-online.target until a network
6162 connection has been configured. The tool primarily integrates
6163 with networkd, but will also make a best effort to make sense
6164 of network configuration performed in some other way.
6166 * Two new service options StartupCPUShares= and
6167 StartupBlockIOWeight= have been added that work similarly to
6168 CPUShares= and BlockIOWeight= however only apply during
6169 system startup. This is useful to prioritize certain services
6170 differently during bootup than during normal runtime.
6172 * hostnamed has been changed to prefer the statically
6173 configured hostname in /etc/hostname (unless set to
6174 'localhost' or empty) over any dynamic one supplied by
6175 dhcp. With this change, the rules for picking the hostname
6176 match more closely the rules of other configuration settings
6177 where the local administrator's configuration in /etc always
6178 overrides any other settings.
6180 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
6181 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
6182 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
6183 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
6184 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
6185 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
6186 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
6187 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
6188 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
6189 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
6190 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
6191 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
6192 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
6193 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
6194 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
6195 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6198 — Beijing, 2014-05-28
6202 * When restoring the screen brightness at boot, stay away from
6203 the darkest setting or from the lowest 5% of the available
6204 range, depending on which is the larger value of both. This
6205 should effectively protect the user from rebooting into a
6206 black screen, should the brightness have been set to minimum
6209 * sd-login gained a new sd_machine_get_class() call to
6210 determine the class ("vm" or "container") of a machine
6211 registered with machined.
6213 * sd-login gained new calls
6214 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
6215 to query the identity of the peer of a local AF_UNIX
6216 connection. They operate similarly to their sd_pid_get_xyz()
6219 * PID 1 will now maintain a system-wide system state engine
6220 with the states "starting", "running", "degraded",
6221 "maintenance", "stopping". These states are bound to system
6222 startup, normal runtime, runtime with at least one failed
6223 service, rescue/emergency mode and system shutdown. This
6224 state is shown in the "systemctl status" output when no unit
6225 name is passed. It is useful to determine system state, in
6226 particularly when doing so for many systems or containers at
6229 * A new command "list-machines" has been added to "systemctl"
6230 that lists all local OS containers and shows their system
6231 state (see above), if systemd runs inside of them.
6233 * systemctl gained a new "-r" switch to recursively enumerate
6234 units on all local containers, when used with the
6235 "list-unit" command (which is the default one that is
6236 executed when no parameters are specified).
6238 * The GPT automatic partition discovery logic will now honour
6239 two GPT partition flags: one may be set on a partition to
6240 cause it to be mounted read-only, and the other may be set
6241 on a partition to ignore it during automatic discovery.
6243 * Two new GPT type UUIDs have been added for automatic root
6244 partition discovery, for 32-bit and 64-bit ARM. This is not
6245 particularly useful for discovering the root directory on
6246 these architectures during bare-metal boots (since UEFI is
6247 not common there), but still very useful to allow booting of
6248 ARM disk images in nspawn with the -i option.
6250 * MAC addresses of interfaces created with nspawn's
6251 --network-interface= switch will now be generated from the
6252 machine name, and thus be stable between multiple invocations
6255 * logind will now automatically remove all IPC objects owned
6256 by a user if she or he fully logs out. This makes sure that
6257 users who are logged out cannot continue to consume IPC
6258 resources. This covers SysV memory, semaphores and message
6259 queues as well as POSIX shared memory and message
6260 queues. Traditionally, SysV and POSIX IPC had no lifecycle
6261 limits. With this functionality, that is corrected. This may
6262 be turned off by using the RemoveIPC= switch of logind.conf.
6264 * The systemd-machine-id-setup and tmpfiles tools gained a
6265 --root= switch to operate on a specific root directory,
6268 * journald can now forward logged messages to the TTYs of all
6269 logged in users ("wall"). This is the default for all
6270 emergency messages now.
6272 * A new tool systemd-journal-remote has been added to stream
6273 journal log messages across the network.
6275 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
6276 controller trees are mounted into it. Note that the
6277 directories mounted beneath it are not read-only. This is a
6278 security measure and is particularly useful because glibc
6279 actually includes a search logic to pick any tmpfs it can
6280 find to implement shm_open() if /dev/shm is not available
6281 (which it might very well be in namespaced setups).
6283 * machinectl gained a new "poweroff" command to cleanly power
6284 down a local OS container.
6286 * The PrivateDevices= unit file setting will now also drop the
6287 CAP_MKNOD capability from the capability bound set, and
6288 imply DevicePolicy=closed.
6290 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
6291 comprehensively on all long-running systemd services where
6292 this is appropriate.
6294 * systemd-udevd will now run in a disassociated mount
6295 namespace. To mount directories from udev rules, make sure to
6296 pull in mount units via SYSTEMD_WANTS properties.
6298 * The kdbus support gained support for uploading policy into
6299 the kernel. sd-bus gained support for creating "monitoring"
6300 connections that can eavesdrop into all bus communication
6301 for debugging purposes.
6303 * Timestamps may now be specified in seconds since the UNIX
6304 epoch Jan 1st, 1970 by specifying "@" followed by the value
6307 * Native tcpwrap support in systemd has been removed. tcpwrap
6308 is old code, not really maintained anymore and has serious
6309 shortcomings, and better options such as firewalls
6310 exist. For setups that require tcpwrap usage, please
6311 consider invoking your socket-activated service via tcpd,
6312 like on traditional inetd.
6314 * A new system.conf configuration option
6315 DefaultTimerAccuracySec= has been added that controls the
6316 default AccuracySec= setting of .timer units.
6318 * Timer units gained a new WakeSystem= switch. If enabled,
6319 timers configured this way will cause the system to resume
6320 from system suspend (if the system supports that, which most
6323 * Timer units gained a new Persistent= switch. If enabled,
6324 timers configured this way will save to disk when they have
6325 been last triggered. This information is then used on next
6326 reboot to possible execute overdue timer events, that
6327 could not take place because the system was powered off.
6328 This enables simple anacron-like behaviour for timer units.
6330 * systemctl's "list-timers" will now also list the time a
6331 timer unit was last triggered in addition to the next time
6332 it will be triggered.
6334 * systemd-networkd will now assign predictable IPv4LL
6335 addresses to its local interfaces.
6337 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
6338 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
6339 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
6340 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
6341 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
6342 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
6343 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
6344 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
6347 — Berlin, 2014-03-25
6351 * A new unit file setting RestrictAddressFamilies= has been
6352 added to restrict which socket address families unit
6353 processes gain access to. This takes address family names
6354 like "AF_INET" or "AF_UNIX", and is useful to minimize the
6355 attack surface of services via exotic protocol stacks. This
6356 is built on seccomp system call filters.
6358 * Two new unit file settings RuntimeDirectory= and
6359 RuntimeDirectoryMode= have been added that may be used to
6360 manage a per-daemon runtime directories below /run. This is
6361 an alternative for setting up directory permissions with
6362 tmpfiles snippets, and has the advantage that the runtime
6363 directory's lifetime is bound to the daemon runtime and that
6364 the daemon starts up with an empty directory each time. This
6365 is particularly useful when writing services that drop
6366 privileges using the User= or Group= setting.
6368 * The DeviceAllow= unit setting now supports globbing for
6369 matching against device group names.
6371 * The systemd configuration file system.conf gained new
6372 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
6373 DefaultMemoryAccounting= to globally turn on/off accounting
6374 for specific resources (cgroups) for all units. These
6375 settings may still be overridden individually in each unit
6378 * systemd-gpt-auto-generator is now able to discover /srv and
6379 root partitions in addition to /home and swap partitions. It
6380 also supports LUKS-encrypted partitions now. With this in
6381 place, automatic discovery of partitions to mount following
6382 the Discoverable Partitions Specification
6383 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
6384 is now a lot more complete. This allows booting without
6385 /etc/fstab and without root= on the kernel command line on
6386 systems prepared appropriately.
6388 * systemd-nspawn gained a new --image= switch which allows
6389 booting up disk images and Linux installations on any block
6390 device that follow the Discoverable Partitions Specification
6391 (see above). This means that installations made with
6392 appropriately updated installers may now be started and
6393 deployed using container managers, completely
6394 unmodified. (We hope that libvirt-lxc will add support for
6395 this feature soon, too.)
6397 * systemd-nspawn gained a new --network-macvlan= setting to
6398 set up a private macvlan interface for the
6399 container. Similarly, systemd-networkd gained a new
6400 Kind=macvlan setting in .netdev files.
6402 * systemd-networkd now supports configuring local addresses
6405 * A new tool systemd-network-wait-online has been added to
6406 synchronously wait for network connectivity using
6409 * The sd-bus.h bus API gained a new sd_bus_track object for
6410 tracking the lifecycle of bus peers. Note that sd-bus.h is
6411 still not a public API though (unless you specify
6412 --enable-kdbus on the configure command line, which however
6413 voids your warranty and you get no API stability guarantee).
6415 * The $XDG_RUNTIME_DIR runtime directories for each user are
6416 now individual tmpfs instances, which has the benefit of
6417 introducing separate pools for each user, with individual
6418 size limits, and thus making sure that unprivileged clients
6419 can no longer negatively impact the system or other users by
6420 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
6421 RuntimeDirectorySize= has been introduced that allows
6422 controlling the default size limit for all users. It
6423 defaults to 10% of the available physical memory. This is no
6424 replacement for quotas on tmpfs though (which the kernel
6425 still does not support), as /dev/shm and /tmp are still
6426 shared resources used by both the system and unprivileged
6429 * logind will now automatically turn off automatic suspending
6430 on laptop lid close when more than one display is
6431 connected. This was previously expected to be implemented
6432 individually in desktop environments (such as GNOME),
6433 however has been added to logind now, in order to fix a
6434 boot-time race where a desktop environment might not have
6435 been started yet and thus not been able to take an inhibitor
6436 lock at the time where logind already suspends the system
6437 due to a closed lid.
6439 * logind will now wait at least 30s after each system
6440 suspend/resume cycle, and 3min after system boot before
6441 suspending the system due to a closed laptop lid. This
6442 should give USB docking stations and similar enough time to
6443 be probed and configured after system resume and boot in
6444 order to then act as suspend blocker.
6446 * systemd-run gained a new --property= setting which allows
6447 initialization of resource control properties (and others)
6448 for the created scope or service unit. Example: "systemd-run
6449 --property=BlockIOWeight=10 updatedb" may be used to run
6450 updatedb at a low block IO scheduling weight.
6452 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
6453 now also work in --scope mode.
6455 * When systemd is compiled with kdbus support, basic support
6456 for enforced policies is now in place. (Note that enabling
6457 kdbus still voids your warranty and no API compatibility
6460 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
6461 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
6462 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
6463 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
6464 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
6465 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
6466 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
6467 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
6468 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
6469 Zbigniew Jędrzejewski-Szmek
6471 — Berlin, 2014-03-12
6475 * systemd will now relabel /dev after loading the SMACK policy
6476 according to SMACK rules.
6478 * A new unit file option AppArmorProfile= has been added to
6479 set the AppArmor profile for the processes of a unit.
6481 * A new condition check ConditionArchitecture= has been added
6482 to conditionalize units based on the system architecture, as
6483 reported by uname()'s "machine" field.
6485 * systemd-networkd now supports matching on the system
6486 virtualization, architecture, kernel command line, host name
6489 * logind is now a lot more aggressive when suspending the
6490 machine due to a closed laptop lid. Instead of acting only
6491 on the lid close action, it will continuously watch the lid
6492 status and act on it. This is useful for laptops where the
6493 power button is on the outside of the chassis so that it can
6494 be reached without opening the lid (such as the Lenovo
6495 Yoga). On those machines, logind will now immediately
6496 re-suspend the machine if the power button has been
6497 accidentally pressed while the laptop was suspended and in a
6498 backpack or similar.
6500 * logind will now watch SW_DOCK switches and inhibit reaction
6501 to the lid switch if it is pressed. This means that logind
6502 will not suspend the machine anymore if the lid is closed
6503 and the system is docked, if the laptop supports SW_DOCK
6504 notifications via the input layer. Note that ACPI docking
6505 stations do not generate this currently. Also note that this
6506 logic is usually not fully sufficient and Desktop
6507 Environments should take a lid switch inhibitor lock when an
6508 external display is connected, as systemd will not watch
6511 * nspawn will now make use of the devices cgroup controller by
6512 default, and only permit creation of and access to the usual
6513 API device nodes like /dev/null or /dev/random, as well as
6514 access to (but not creation of) the pty devices.
6516 * We will now ship a default .network file for
6517 systemd-networkd that automatically configures DHCP for
6518 network interfaces created by nspawn's --network-veth or
6519 --network-bridge= switches.
6521 * systemd will now understand the usual M, K, G, T suffixes
6522 according to SI conventions (i.e. to the base 1000) when
6523 referring to throughput and hardware metrics. It will stay
6524 with IEC conventions (i.e. to the base 1024) for software
6525 metrics, according to what is customary according to
6526 Wikipedia. We explicitly document which base applies for
6527 each configuration option.
6529 * The DeviceAllow= setting in unit files now supports a syntax
6530 to whitelist an entire group of devices node majors at once,
6531 based on the /proc/devices listing. For example, with the
6532 string "char-pts", it is now possible to whitelist all
6533 current and future pseudo-TTYs at once.
6535 * sd-event learned a new "post" event source. Event sources of
6536 this type are triggered by the dispatching of any event
6537 source of a type that is not "post". This is useful for
6538 implementing clean-up and check event sources that are
6539 triggered by other work being done in the program.
6541 * systemd-networkd is no longer statically enabled, but uses
6542 the usual [Install] sections so that it can be
6543 enabled/disabled using systemctl. It still is enabled by
6546 * When creating a veth interface pair with systemd-nspawn, the
6547 host side will now be prefixed with "vb-" if
6548 --network-bridge= is used, and with "ve-" if --network-veth
6549 is used. This way, it is easy to distinguish these cases on
6550 the host, for example to apply different configuration to
6551 them with systemd-networkd.
6553 * The compatibility libraries for libsystemd-journal.so,
6554 libsystem-id128.so, libsystemd-login.so and
6555 libsystemd-daemon.so do not make use of IFUNC
6556 anymore. Instead, we now build libsystemd.so multiple times
6557 under these alternative names. This means that the footprint
6558 is drastically increased, but given that these are
6559 transitional compatibility libraries, this should not matter
6560 much. This change has been made necessary to support the ARM
6561 platform for these compatibility libraries, as the ARM
6562 toolchain is not really at the same level as the toolchain
6563 for other architectures like x86 and does not support
6564 IFUNC. Please make sure to use --enable-compat-libs only
6565 during a transitional period!
6567 * The .include syntax has been deprecated and is not documented
6568 anymore. Drop-in files in .d directories should be used instead.
6570 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
6571 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
6572 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
6573 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
6574 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
6575 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
6576 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
6577 Zbigniew Jędrzejewski-Szmek
6579 — Berlin, 2014-02-24
6583 * A new component "systemd-networkd" has been added that can
6584 be used to configure local network interfaces statically or
6585 via DHCP. It is capable of bringing up bridges, VLANs, and
6586 bonding. Currently, no hook-ups for interactive network
6587 configuration are provided. Use this for your initrd,
6588 container, embedded, or server setup if you need a simple,
6589 yet powerful, network configuration solution. This
6590 configuration subsystem is quite nifty, as it allows wildcard
6591 hotplug matching in interfaces. For example, with a single
6592 configuration snippet, you can configure that all Ethernet
6593 interfaces showing up are automatically added to a bridge,
6594 or similar. It supports link-sensing and more.
6596 * A new tool "systemd-socket-proxyd" has been added which can
6597 act as a bidirectional proxy for TCP sockets. This is
6598 useful for adding socket activation support to services that
6599 do not actually support socket activation, including virtual
6600 machines and the like.
6602 * Add a new tool to save/restore rfkill state on
6605 * Save/restore state of keyboard backlights in addition to
6606 display backlights on shutdown/boot.
6608 * udev learned a new SECLABEL{} construct to label device
6609 nodes with a specific security label when they appear. For
6610 now, only SECLABEL{selinux} is supported, but the syntax is
6611 prepared for additional security frameworks.
6613 * udev gained a new scheme to configure link-level attributes
6614 from files in /etc/systemd/network/*.link. These files can
6615 match against MAC address, device path, driver name and type,
6616 and will apply attributes like the naming policy, link speed,
6617 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
6618 address assignment policy (randomized, ...).
6620 * The configuration of network interface naming rules for
6621 "permanent interface names" has changed: a new NamePolicy=
6622 setting in the [Link] section of .link files determines the
6623 priority of possible naming schemes (onboard, slot, MAC,
6624 path). The default value of this setting is determined by
6625 /usr/lib/net/links/99-default.link. Old
6626 80-net-name-slot.rules udev configuration file has been
6627 removed, so local configuration overriding this file should
6628 be adapted to override 99-default.link instead.
6630 * When the User= switch is used in a unit file, also
6631 initialize $SHELL= based on the user database entry.
6633 * systemd no longer depends on libdbus. All communication is
6634 now done with sd-bus, systemd's low-level bus library
6637 * kdbus support has been added to PID 1 itself. When kdbus is
6638 enabled, this causes PID 1 to set up the system bus and
6639 enable support for a new ".busname" unit type that
6640 encapsulates bus name activation on kdbus. It works a little
6641 bit like ".socket" units, except for bus names. A new
6642 generator has been added that converts classic dbus1 service
6643 activation files automatically into native systemd .busname
6646 * sd-bus: add a light-weight vtable implementation that allows
6647 defining objects on the bus with a simple static const
6648 vtable array of its methods, signals and properties.
6650 * systemd will not generate or install static dbus
6651 introspection data anymore to /usr/share/dbus-1/interfaces,
6652 as the precise format of these files is unclear, and
6653 nothing makes use of it.
6655 * A proxy daemon is now provided to proxy clients connecting
6656 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
6657 compatibility with classic D-Bus.
6659 * A bus driver implementation has been added that supports the
6660 classic D-Bus bus driver calls on kdbus, also for
6661 compatibility purposes.
6663 * A new API "sd-event.h" has been added that implements a
6664 minimal event loop API built around epoll. It provides a
6665 couple of features that direct epoll usage is lacking:
6666 prioritization of events, scales to large numbers of timer
6667 events, per-event timer slack (accuracy), system-wide
6668 coalescing of timer events, exit handlers, watchdog
6669 supervision support using systemd's sd_notify() API, child
6672 * A new API "sd-rntl.h" has been added that provides an API
6673 around the route netlink interface of the kernel, similar in
6674 style to "sd-bus.h".
6676 * A new API "sd-dhcp-client.h" has been added that provides a
6677 small DHCPv4 client-side implementation. This is used by
6680 * There is a new kernel command line option
6681 "systemd.restore_state=0|1". When set to "0", none of the
6682 systemd tools will restore saved runtime state to hardware
6683 devices. More specifically, the rfkill and backlight states
6686 * The FsckPassNo= compatibility option in mount/service units
6687 has been removed. The fstab generator will now add the
6688 necessary dependencies automatically, and does not require
6689 PID1's support for that anymore.
6691 * journalctl gained a new switch, --list-boots, that lists
6692 recent boots with their times and boot IDs.
6694 * The various tools like systemctl, loginctl, timedatectl,
6695 busctl, systemd-run, ... have gained a new switch "-M" to
6696 connect to a specific, local OS container (as direct
6697 connection, without requiring SSH). This works on any
6698 container that is registered with machined, such as those
6699 created by libvirt-lxc or nspawn.
6701 * systemd-run and systemd-analyze also gained support for "-H"
6702 to connect to remote hosts via SSH. This is particularly
6703 useful for systemd-run because it enables queuing of jobs
6704 onto remote systems.
6706 * machinectl gained a new command "login" to open a getty
6707 login in any local container. This works with any container
6708 that is registered with machined (such as those created by
6709 libvirt-lxc or nspawn), and which runs systemd inside.
6711 * machinectl gained a new "reboot" command that may be used to
6712 trigger a reboot on a specific container that is registered
6713 with machined. This works on any container that runs an init
6714 system of some kind.
6716 * systemctl gained a new "list-timers" command to print a nice
6717 listing of installed timer units with the times they elapse
6720 * Alternative reboot() parameters may now be specified on the
6721 "systemctl reboot" command line and are passed to the
6722 reboot() system call.
6724 * systemctl gained a new --job-mode= switch to configure the
6725 mode to queue a job with. This is a more generic version of
6726 --fail, --irreversible, and --ignore-dependencies, which are
6727 still available but not advertised anymore.
6729 * /etc/systemd/system.conf gained new settings to configure
6730 various default timeouts of units, as well as the default
6731 start limit interval and burst. These may still be overridden
6734 * PID1 will now export on the bus profile data of the security
6735 policy upload process (such as the SELinux policy upload to
6738 * journald: when forwarding logs to the console, include
6739 timestamps (following the setting in
6740 /sys/module/printk/parameters/time).
6742 * OnCalendar= in timer units now understands the special
6743 strings "yearly" and "annually". (Both are equivalent)
6745 * The accuracy of timer units is now configurable with the new
6746 AccuracySec= setting. It defaults to 1min.
6748 * A new dependency type JoinsNamespaceOf= has been added that
6749 allows running two services within the same /tmp and network
6750 namespace, if PrivateNetwork= or PrivateTmp= are used.
6752 * A new command "cat" has been added to systemctl. It outputs
6753 the original unit file of a unit, and concatenates the
6754 contents of additional "drop-in" unit file snippets, so that
6755 the full configuration is shown.
6757 * systemctl now supports globbing on the various "list-xyz"
6758 commands, like "list-units" or "list-sockets", as well as on
6759 those commands which take multiple unit names.
6761 * journalctl's --unit= switch gained support for globbing.
6763 * All systemd daemons now make use of the watchdog logic so
6764 that systemd automatically notices when they hang.
6766 * If the $container_ttys environment variable is set,
6767 getty-generator will automatically spawn a getty for each
6768 listed tty. This is useful for container managers to request
6769 login gettys to be spawned on as many ttys as needed.
6771 * %h, %s, %U specifier support is not available anymore when
6772 used in unit files for PID 1. This is because NSS calls are
6773 not safe from PID 1. They stay available for --user
6774 instances of systemd, and as special case for the root user.
6776 * loginctl gained a new "--no-legend" switch to turn off output
6779 * The "sd-login.h" API gained three new calls:
6780 sd_session_is_remote(), sd_session_get_remote_user(),
6781 sd_session_get_remote_host() to query information about
6784 * The udev hardware database now also carries vendor/product
6785 information of SDIO devices.
6787 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
6788 determine whether watchdog notifications are requested by
6791 * Socket-activated per-connection services now include a
6792 short description of the connection parameters in the
6795 * tmpfiles gained a new "--boot" option. When this is not used,
6796 only lines where the command character is not suffixed with
6797 "!" are executed. When this option is specified, those
6798 options are executed too. This partitions tmpfiles
6799 directives into those that can be safely executed at any
6800 time, and those which should be run only at boot (for
6801 example, a line that creates /run/nologin).
6803 * A new API "sd-resolve.h" has been added which provides a simple
6804 asynchronous wrapper around glibc NSS host name resolution
6805 calls, such as getaddrinfo(). In contrast to glibc's
6806 getaddrinfo_a(), it does not use signals. In contrast to most
6807 other asynchronous name resolution libraries, this one does
6808 not reimplement DNS, but reuses NSS, so that alternate
6809 host name resolution systems continue to work, such as mDNS,
6810 LDAP, etc. This API is based on libasyncns, but it has been
6811 cleaned up for inclusion in systemd.
6813 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
6814 "sd-daemon.h" are no longer found in individual libraries
6815 libsystemd-journal.so, libsystemd-login.so,
6816 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
6817 merged them into a single library, libsystemd.so, which
6818 provides all symbols. The reason for this is cyclic
6819 dependencies, as these libraries tend to use each other's
6820 symbols. So far, we have managed to workaround that by linking
6821 a copy of a good part of our code into each of these
6822 libraries again and again, which, however, makes certain
6823 things hard to do, like sharing static variables. Also, it
6824 substantially increases footprint. With this change, there
6825 is only one library for the basic APIs systemd
6826 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
6827 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
6828 library as well, however are subject to the --enable-kdbus
6829 switch (see below). Note that "sd-dhcp-client.h" is not part
6830 of this library (this is because it only consumes, never
6831 provides, services of/to other APIs). To make the transition
6832 easy from the separate libraries to the unified one, we
6833 provide the --enable-compat-libs compile-time switch which
6834 will generate stub libraries that are compatible with the
6835 old ones but redirect all calls to the new one.
6837 * All of the kdbus logic and the new APIs "sd-bus.h",
6838 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
6839 and "sd-utf8.h" are compile-time optional via the
6840 "--enable-kdbus" switch, and they are not compiled in by
6841 default. To make use of kdbus, you have to explicitly enable
6842 the switch. Note however, that neither the kernel nor the
6843 userspace API for all of this is considered stable yet. We
6844 want to maintain the freedom to still change the APIs for
6845 now. By specifying this build-time switch, you acknowledge
6846 that you are aware of the instability of the current
6849 * Also, note that while kdbus is pretty much complete,
6850 it lacks one thing: proper policy support. This means you
6851 can build a fully working system with all features; however,
6852 it will be highly insecure. Policy support will be added in
6853 one of the next releases, at the same time that we will
6854 declare the APIs stable.
6856 * When the kernel command line argument "kdbus" is specified,
6857 systemd will automatically load the kdbus.ko kernel module. At
6858 this stage of development, it is only useful for testing kdbus
6859 and should not be used in production. Note: if "--enable-kdbus"
6860 is specified, and the kdbus.ko kernel module is available, and
6861 "kdbus" is added to the kernel command line, the entire system
6862 runs with kdbus instead of dbus-daemon, with the above mentioned
6863 problem of missing the system policy enforcement. Also a future
6864 version of kdbus.ko or a newer systemd will not be compatible with
6865 each other, and will unlikely be able to boot the machine if only
6866 one of them is updated.
6868 * systemctl gained a new "import-environment" command which
6869 uploads the caller's environment (or parts thereof) into the
6870 service manager so that it is inherited by services started
6871 by the manager. This is useful to upload variables like
6872 $DISPLAY into the user service manager.
6874 * A new PrivateDevices= switch has been added to service units
6875 which allows running a service with a namespaced /dev
6876 directory that does not contain any device nodes for
6877 physical devices. More specifically, it only includes devices
6878 such as /dev/null, /dev/urandom, and /dev/zero which are API
6881 * logind has been extended to support behaviour like VT
6882 switching on seats that do not support a VT. This makes
6883 multi-session available on seats that are not the first seat
6884 (seat0), and on systems where kernel support for VTs has
6885 been disabled at compile-time.
6887 * If a process holds a delay lock for system sleep or shutdown
6888 and fails to release it in time, we will now log its
6889 identity. This makes it easier to identify processes that
6890 cause slow suspends or power-offs.
6892 * When parsing /etc/crypttab, support for a new key-slot=
6893 option as supported by Debian is added. It allows indicating
6894 which LUKS slot to use on disk, speeding up key loading.
6896 * The sd_journal_sendv() API call has been checked and
6897 officially declared to be async-signal-safe so that it may
6898 be invoked from signal handlers for logging purposes.
6900 * Boot-time status output is now enabled automatically after a
6901 short timeout if boot does not progress, in order to give
6902 the user an indication what she or he is waiting for.
6904 * The boot-time output has been improved to show how much time
6905 remains until jobs expire.
6907 * The KillMode= switch in service units gained a new possible
6908 value "mixed". If set, and the unit is shut down, then the
6909 initial SIGTERM signal is sent only to the main daemon
6910 process, while the following SIGKILL signal is sent to
6911 all remaining processes of the service.
6913 * When a scope unit is registered, a new property "Controller"
6914 may be set. If set to a valid bus name, systemd will send a
6915 RequestStop() signal to this name when it would like to shut
6916 down the scope. This may be used to hook manager logic into
6917 the shutdown logic of scope units. Also, scope units may now
6918 be put in a special "abandoned" state, in which case the
6919 manager process which created them takes no further
6920 responsibilities for it.
6922 * When reading unit files, systemd will now verify
6923 the access mode of these files, and warn about certain
6924 suspicious combinations. This has been added to make it
6925 easier to track down packaging bugs where unit files are
6926 marked executable or world-writable.
6928 * systemd-nspawn gained a new "--setenv=" switch to set
6929 container-wide environment variables. The similar option in
6930 systemd-activate was renamed from "--environment=" to
6931 "--setenv=" for consistency.
6933 * systemd-nspawn has been updated to create a new kdbus domain
6934 for each container that is invoked, thus allowing each
6935 container to have its own set of system and user buses,
6936 independent of the host.
6938 * systemd-nspawn gained a new --drop-capability= switch to run
6939 the container with less capabilities than the default. Both
6940 --drop-capability= and --capability= now take the special
6941 string "all" for dropping or keeping all capabilities.
6943 * systemd-nspawn gained new switches for executing containers
6944 with specific SELinux labels set.
6946 * systemd-nspawn gained a new --quiet switch to not generate
6947 any additional output but the container's own console
6950 * systemd-nspawn gained a new --share-system switch to run a
6951 container without PID namespacing enabled.
6953 * systemd-nspawn gained a new --register= switch to control
6954 whether the container is registered with systemd-machined or
6955 not. This is useful for containers that do not run full
6956 OS images, but only specific apps.
6958 * systemd-nspawn gained a new --keep-unit which may be used
6959 when invoked as the only program from a service unit, and
6960 results in registration of the unit service itself in
6961 systemd-machined, instead of a newly opened scope unit.
6963 * systemd-nspawn gained a new --network-interface= switch for
6964 moving arbitrary interfaces to the container. The new
6965 --network-veth switch creates a virtual Ethernet connection
6966 between host and container. The new --network-bridge=
6967 switch then allows assigning the host side of this virtual
6968 Ethernet connection to a bridge device.
6970 * systemd-nspawn gained a new --personality= switch for
6971 setting the kernel personality for the container. This is
6972 useful when running a 32-bit container on a 64-bit host. A
6973 similar option Personality= is now also available for service
6976 * logind will now also track a "Desktop" identifier for each
6977 session which encodes the desktop environment of it. This is
6978 useful for desktop environments that want to identify
6979 multiple running sessions of itself easily.
6981 * A new SELinuxContext= setting for service units has been
6982 added that allows setting a specific SELinux execution
6983 context for a service.
6985 * Most systemd client tools will now honour $SYSTEMD_LESS for
6986 settings of the "less" pager. By default, these tools will
6987 override $LESS to allow certain operations to work, such as
6988 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
6989 influence this logic.
6991 * systemd's "seccomp" hook-up has been changed to make use of
6992 the libseccomp library instead of using its own
6993 implementation. This has benefits for portability among
6996 * For usage together with SystemCallFilter=, a new
6997 SystemCallErrorNumber= setting has been introduced that
6998 allows configuration of a system error number to be returned
6999 on filtered system calls, instead of immediately killing the
7000 process. Also, SystemCallArchitectures= has been added to
7001 limit access to system calls of a particular architecture
7002 (in order to turn off support for unused secondary
7003 architectures). There is also a global
7004 SystemCallArchitectures= setting in system.conf now to turn
7005 off support for non-native system calls system-wide.
7007 * systemd requires a kernel with a working name_to_handle_at(),
7008 please see the kernel config requirements in the README file.
7010 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
7011 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
7012 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
7013 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
7014 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
7015 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
7016 Elia Pinto, Florian Weimer, George McCollister, Goffredo
7017 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
7018 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
7019 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
7020 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
7021 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
7022 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
7023 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
7024 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
7025 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
7026 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
7027 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
7028 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
7029 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
7030 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
7031 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
7032 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
7033 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
7035 — Berlin, 2014-02-20
7039 * logind has gained support for facilitating privileged input
7040 and drm device access for unprivileged clients. This work is
7041 useful to allow Wayland display servers (and similar
7042 programs, such as kmscon) to run under the user's ID and
7043 access input and drm devices which are normally
7044 protected. When this is used (and the kernel is new enough)
7045 logind will "mute" IO on the file descriptors passed to
7046 Wayland as long as it is in the background and "unmute" it
7047 if it returns into the foreground. This allows secure
7048 session switching without allowing background sessions to
7049 eavesdrop on input and display data. This also introduces
7050 session switching support if VT support is turned off in the
7051 kernel, and on seats that are not seat0.
7053 * A new kernel command line option luks.options= is understood
7054 now which allows specifying LUKS options for usage for LUKS
7055 encrypted partitions specified with luks.uuid=.
7057 * tmpfiles.d(5) snippets may now use specifier expansion in
7058 path names. More specifically %m, %b, %H, %v, are now
7059 replaced by the local machine id, boot id, hostname, and
7060 kernel version number.
7062 * A new tmpfiles.d(5) command "m" has been introduced which
7063 may be used to change the owner/group/access mode of a file
7064 or directory if it exists, but do nothing if it does not.
7066 * This release removes high-level support for the
7067 MemorySoftLimit= cgroup setting. The underlying kernel
7068 cgroup attribute memory.soft_limit= is currently badly
7069 designed and likely to be removed from the kernel API in its
7070 current form, hence we should not expose it for now.
7072 * The memory.use_hierarchy cgroup attribute is now enabled for
7073 all cgroups systemd creates in the memory cgroup
7074 hierarchy. This option is likely to be come the built-in
7075 default in the kernel anyway, and the non-hierarchical mode
7076 never made much sense in the intrinsically hierarchical
7079 * A new field _SYSTEMD_SLICE= is logged along with all journal
7080 messages containing the slice a message was generated
7081 from. This is useful to allow easy per-customer filtering of
7082 logs among other things.
7084 * systemd-journald will no longer adjust the group of journal
7085 files it creates to the "systemd-journal" group. Instead we
7086 rely on the journal directory to be owned by the
7087 "systemd-journal" group, and its setgid bit set, so that the
7088 kernel file system layer will automatically enforce that
7089 journal files inherit this group assignment. The reason for
7090 this change is that we cannot allow NSS look-ups from
7091 journald which would be necessary to resolve
7092 "systemd-journal" to a numeric GID, because this might
7093 create deadlocks if NSS involves synchronous queries to
7094 other daemons (such as nscd, or sssd) which in turn are
7095 logging clients of journald and might block on it, which
7096 would then dead lock. A tmpfiles.d(5) snippet included in
7097 systemd will make sure the setgid bit and group are
7098 properly set on the journal directory if it exists on every
7099 boot. However, we recommend adjusting it manually after
7100 upgrades too (or from RPM scriptlets), so that the change is
7101 not delayed until next reboot.
7103 * Backlight and random seed files in /var/lib/ have moved into
7104 the /var/lib/systemd/ directory, in order to centralize all
7105 systemd generated files in one directory.
7107 * Boot time performance measurements (as displayed by
7108 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
7109 performance information if that's available to determine how
7110 much time BIOS and boot loader initialization required. With
7111 a sufficiently new BIOS you hence no longer need to boot
7112 with Gummiboot to get access to such information.
7114 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
7115 Cristian Rodríguez, Dave Reisner, David Herrmann, David
7116 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
7117 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
7118 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
7119 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
7120 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7122 — Berlin, 2013-10-02
7126 * The Restart= option for services now understands a new
7127 on-watchdog setting, which will restart the service
7128 automatically if the service stops sending out watchdog keep
7129 alive messages (as configured with WatchdogSec=).
7131 * The getty generator (which is responsible for bringing up a
7132 getty on configured serial consoles) will no longer only
7133 start a getty on the primary kernel console but on all
7134 others, too. This makes the order in which console= is
7135 specified on the kernel command line less important.
7137 * libsystemd-logind gained a new sd_session_get_vt() call to
7138 retrieve the VT number of a session.
7140 * If the option "tries=0" is set for an entry of /etc/crypttab
7141 its passphrase is queried indefinitely instead of any
7142 maximum number of tries.
7144 * If a service with a configure PID file terminates its PID
7145 file will now be removed automatically if it still exists
7146 afterwards. This should put an end to stale PID files.
7148 * systemd-run will now also take relative binary path names
7149 for execution and no longer insists on absolute paths.
7151 * InaccessibleDirectories= and ReadOnlyDirectories= now take
7152 paths that are optionally prefixed with "-" to indicate that
7153 it should not be considered a failure if they do not exist.
7155 * journalctl -o (and similar commands) now understands a new
7156 output mode "short-precise", it is similar to "short" but
7157 shows timestamps with usec accuracy.
7159 * The option "discard" (as known from Debian) is now
7160 synonymous to "allow-discards" in /etc/crypttab. In fact,
7161 "discard" is preferred now (since it is easier to remember
7164 * Some licensing clean-ups were made, so that more code is now
7165 LGPL-2.1 licensed than before.
7167 * A minimal tool to save/restore the display backlight
7168 brightness across reboots has been added. It will store the
7169 backlight setting as late as possible at shutdown, and
7170 restore it as early as possible during reboot.
7172 * A logic to automatically discover and enable home and swap
7173 partitions on GPT disks has been added. With this in place
7174 /etc/fstab becomes optional for many setups as systemd can
7175 discover certain partitions located on the root disk
7176 automatically. Home partitions are recognized under their
7177 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
7178 partitions are recognized under their GPT type ID
7179 0657fd6da4ab43c484e50933c84b4f4f.
7181 * systemd will no longer pass any environment from the kernel
7182 or initrd to system services. If you want to set an
7183 environment for all services, do so via the kernel command
7184 line systemd.setenv= assignment.
7186 * The systemd-sysctl tool no longer natively reads the file
7187 /etc/sysctl.conf. If desired, the file should be symlinked
7188 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
7189 legacy support by a symlink rather than built-in code, it
7190 also makes the otherwise hidden order of application of the
7191 different files visible. (Note that this partly reverts to a
7192 pre-198 application order of sysctl knobs!)
7194 * The "systemctl set-log-level" and "systemctl dump" commands
7195 have been moved to systemd-analyze.
7197 * systemd-run learned the new --remain-after-exit switch,
7198 which causes the scope unit not to be cleaned up
7199 automatically after the process terminated.
7201 * tmpfiles learned a new --exclude-prefix= switch to exclude
7202 certain paths from operation.
7204 * journald will now automatically flush all messages to disk
7205 as soon as a message at the log level CRIT, ALERT or EMERG
7208 Contributions from: Andrew Cook, Brandon Philips, Christian
7209 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
7210 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
7211 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
7212 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
7213 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
7214 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
7215 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
7216 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
7217 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
7218 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
7219 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
7220 William Giokas, Zbigniew Jędrzejewski-Szmek
7222 — Berlin, 2013-09-13
7226 * The documentation has been updated to cover the various new
7227 concepts introduced with 205.
7229 * Unit files now understand the new %v specifier which
7230 resolves to the kernel version string as returned by "uname
7233 * systemctl now supports filtering the unit list output by
7234 load state, active state and sub state, using the new
7237 * "systemctl status" will now show the results of the
7238 condition checks (like ConditionPathExists= and similar) of
7239 the last start attempts of the unit. They are also logged to
7242 * "journalctl -b" may now be used to look for boot output of a
7243 specific boot. Try "journalctl -b -1" for the previous boot,
7244 but the syntax is substantially more powerful.
7246 * "journalctl --show-cursor" has been added which prints the
7247 cursor string the last shown log line. This may then be used
7248 with the new "journalctl --after-cursor=" switch to continue
7249 browsing logs from that point on.
7251 * "journalctl --force" may now be used to force regeneration
7254 * Creation of "dead" device nodes has been moved from udev
7255 into kmod and tmpfiles. Previously, udev would read the kmod
7256 databases to pre-generate dead device nodes based on meta
7257 information contained in kernel modules, so that these would
7258 be auto-loaded on access rather then at boot. As this
7259 does not really have much to do with the exposing actual
7260 kernel devices to userspace this has always been slightly
7261 alien in the udev codebase. Following the new scheme kmod
7262 will now generate a runtime snippet for tmpfiles from the
7263 module meta information and it now is tmpfiles' job to the
7264 create the nodes. This also allows overriding access and
7265 other parameters for the nodes using the usual tmpfiles
7266 facilities. As side effect this allows us to remove the
7267 CAP_SYS_MKNOD capability bit from udevd entirely.
7269 * logind's device ACLs may now be applied to these "dead"
7270 devices nodes too, thus finally allowing managed access to
7271 devices such as /dev/snd/sequencer without loading the
7272 backing module right-away.
7274 * A new RPM macro has been added that may be used to apply
7275 tmpfiles configuration during package installation.
7277 * systemd-detect-virt and ConditionVirtualization= now can
7278 detect User-Mode-Linux machines (UML).
7280 * journald will now implicitly log the effective capabilities
7281 set of processes in the message metadata.
7283 * systemd-cryptsetup has gained support for TrueCrypt volumes.
7285 * The initrd interface has been simplified (more specifically,
7286 support for passing performance data via environment
7287 variables and fsck results via files in /run has been
7288 removed). These features were non-essential, and are
7289 nowadays available in a much nicer way by having systemd in
7290 the initrd serialize its state and have the hosts systemd
7291 deserialize it again.
7293 * The udev "keymap" data files and tools to apply keyboard
7294 specific mappings of scan to key codes, and force-release
7295 scan code lists have been entirely replaced by a udev
7296 "keyboard" builtin and a hwdb data file.
7298 * systemd will now honour the kernel's "quiet" command line
7299 argument also during late shutdown, resulting in a
7300 completely silent shutdown when used.
7302 * There's now an option to control the SO_REUSEPORT socket
7303 option in .socket units.
7305 * Instance units will now automatically get a per-template
7306 subslice of system.slice unless something else is explicitly
7307 configured. For example, instances of sshd@.service will now
7308 implicitly be placed in system-sshd.slice rather than
7309 system.slice as before.
7311 * Test coverage support may now be enabled at build time.
7313 Contributions from: Dave Reisner, Frederic Crozat, Harald
7314 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
7315 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
7316 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
7317 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
7318 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
7319 Giokas, Zbigniew Jędrzejewski-Szmek
7321 — Berlin, 2013-07-23
7325 * Two new unit types have been introduced:
7327 Scope units are very similar to service units, however, are
7328 created out of pre-existing processes — instead of PID 1
7329 forking off the processes. By using scope units it is
7330 possible for system services and applications to group their
7331 own child processes (worker processes) in a powerful way
7332 which then maybe used to organize them, or kill them
7333 together, or apply resource limits on them.
7335 Slice units may be used to partition system resources in an
7336 hierarchical fashion and then assign other units to them. By
7337 default there are now three slices: system.slice (for all
7338 system services), user.slice (for all user sessions),
7339 machine.slice (for VMs and containers).
7341 Slices and scopes have been introduced primarily in
7342 context of the work to move cgroup handling to a
7343 single-writer scheme, where only PID 1
7344 creates/removes/manages cgroups.
7346 * There's a new concept of "transient" units. In contrast to
7347 normal units these units are created via an API at runtime,
7348 not from configuration from disk. More specifically this
7349 means it is now possible to run arbitrary programs as
7350 independent services, with all execution parameters passed
7351 in via bus APIs rather than read from disk. Transient units
7352 make systemd substantially more dynamic then it ever was,
7353 and useful as a general batch manager.
7355 * logind has been updated to make use of scope and slice units
7356 for managing user sessions. As a user logs in he will get
7357 his own private slice unit, to which all sessions are added
7358 as scope units. We also added support for automatically
7359 adding an instance of user@.service for the user into the
7360 slice. Effectively logind will no longer create cgroup
7361 hierarchies on its own now, it will defer entirely to PID 1
7362 for this by means of scope, service and slice units. Since
7363 user sessions this way become entities managed by PID 1
7364 the output of "systemctl" is now a lot more comprehensive.
7366 * A new mini-daemon "systemd-machined" has been added which
7367 may be used by virtualization managers to register local
7368 VMs/containers. nspawn has been updated accordingly, and
7369 libvirt will be updated shortly. machined will collect a bit
7370 of meta information about the VMs/containers, and assign
7371 them their own scope unit (see above). The collected
7372 meta-data is then made available via the "machinectl" tool,
7373 and exposed in "ps" and similar tools. machined/machinectl
7374 is compile-time optional.
7376 * As discussed earlier, the low-level cgroup configuration
7377 options ControlGroup=, ControlGroupModify=,
7378 ControlGroupPersistent=, ControlGroupAttribute= have been
7379 removed. Please use high-level attribute settings instead as
7380 well as slice units.
7382 * A new bus call SetUnitProperties() has been added to alter
7383 various runtime parameters of a unit. This is primarily
7384 useful to alter cgroup parameters dynamically in a nice way,
7385 but will be extended later on to make more properties
7386 modifiable at runtime. systemctl gained a new set-properties
7387 command that wraps this call.
7389 * A new tool "systemd-run" has been added which can be used to
7390 run arbitrary command lines as transient services or scopes,
7391 while configuring a number of settings via the command
7392 line. This tool is currently very basic, however already
7393 very useful. We plan to extend this tool to even allow
7394 queuing of execution jobs with time triggers from the
7395 command line, similar in fashion to "at".
7397 * nspawn will now inform the user explicitly that kernels with
7398 audit enabled break containers, and suggest the user to turn
7401 * Support for detecting the IMA and AppArmor security
7402 frameworks with ConditionSecurity= has been added.
7404 * journalctl gained a new "-k" switch for showing only kernel
7405 messages, mimicking dmesg output; in addition to "--user"
7406 and "--system" switches for showing only user's own logs
7409 * systemd-delta can now show information about drop-in
7410 snippets extending unit files.
7412 * libsystemd-bus has been substantially updated but is still
7413 not available as public API.
7415 * systemd will now look for the "debug" argument on the kernel
7416 command line and enable debug logging, similar to what
7417 "systemd.log_level=debug" already did before.
7419 * "systemctl set-default", "systemctl get-default" has been
7420 added to configure the default.target symlink, which
7421 controls what to boot into by default.
7423 * "systemctl set-log-level" has been added as a convenient
7424 way to raise and lower systemd logging threshold.
7426 * "systemd-analyze plot" will now show the time the various
7427 generators needed for execution, as well as information
7428 about the unit file loading.
7430 * libsystemd-journal gained a new sd_journal_open_files() call
7431 for opening specific journal files. journactl also gained a
7432 new switch to expose this new functionality. Previously we
7433 only supported opening all files from a directory, or all
7434 files from the system, as opening individual files only is
7435 racy due to journal file rotation.
7437 * systemd gained the new DefaultEnvironment= setting in
7438 /etc/systemd/system.conf to set environment variables for
7441 * If a privileged process logs a journal message with the
7442 OBJECT_PID= field set, then journald will automatically
7443 augment this with additional OBJECT_UID=, OBJECT_GID=,
7444 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
7445 system services want to log events about specific client
7446 processes. journactl/systemctl has been updated to make use
7447 of this information if all log messages regarding a specific
7450 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
7451 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
7452 Reisner, David Coppa, David King, David Strauss, Eelco
7453 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
7454 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
7455 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
7456 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
7457 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
7458 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
7459 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
7460 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
7461 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
7462 Łukasz Stelmach, 장동준
7466 * The Python bindings gained some minimal support for the APIs
7467 exposed by libsystemd-logind.
7469 * ConditionSecurity= gained support for detecting SMACK. Since
7470 this condition already supports SELinux and AppArmor we only
7471 miss IMA for this. Patches welcome!
7473 Contributions from: Karol Lewandowski, Lennart Poettering,
7474 Zbigniew Jędrzejewski-Szmek
7478 * systemd-nspawn will now create /etc/resolv.conf if
7479 necessary, before bind-mounting the host's file onto it.
7481 * systemd-nspawn will now store meta information about a
7482 container on the container's cgroup as extended attribute
7483 fields, including the root directory.
7485 * The cgroup hierarchy has been reworked in many ways. All
7486 objects any of the components systemd creates in the cgroup
7487 tree are now suffixed. More specifically, user sessions are
7488 now placed in cgroups suffixed with ".session", users in
7489 cgroups suffixed with ".user", and nspawn containers in
7490 cgroups suffixed with ".nspawn". Furthermore, all cgroup
7491 names are now escaped in a simple scheme to avoid collision
7492 of userspace object names with kernel filenames. This work
7493 is preparation for making these objects relocatable in the
7494 cgroup tree, in order to allow easy resource partitioning of
7495 these objects without causing naming conflicts.
7497 * systemctl list-dependencies gained the new switches
7498 --plain, --reverse, --after and --before.
7500 * systemd-inhibit now shows the process name of processes that
7501 have taken an inhibitor lock.
7503 * nss-myhostname will now also resolve "localhost"
7504 implicitly. This makes /etc/hosts an optional file and
7505 nicely handles that on IPv6 ::1 maps to both "localhost" and
7508 * libsystemd-logind.so gained a new call
7509 sd_get_machine_names() to enumerate running containers and
7510 VMs (currently only supported by very new libvirt and
7511 nspawn). sd_login_monitor can now be used to watch
7512 VMs/containers coming and going.
7514 * .include is not allowed recursively anymore, and only in
7515 unit files. Usually it is better to use drop-in snippets in
7516 .d/*.conf anyway, as introduced with systemd 198.
7518 * systemd-analyze gained a new "critical-chain" command that
7519 determines the slowest chain of units run during system
7520 boot-up. It is very useful for tracking down where
7521 optimizing boot time is the most beneficial.
7523 * systemd will no longer allow manipulating service paths in
7524 the name=systemd:/system cgroup tree using ControlGroup= in
7525 units. (But is still fine with it in all other dirs.)
7527 * There's a new systemd-nspawn@.service service file that may
7528 be used to easily run nspawn containers as system
7529 services. With the container's root directory in
7530 /var/lib/container/foobar it is now sufficient to run
7531 "systemctl start systemd-nspawn@foobar.service" to boot it.
7533 * systemd-cgls gained a new parameter "--machine" to list only
7534 the processes within a certain container.
7536 * ConditionSecurity= now can check for "apparmor". We still
7537 are lacking checks for SMACK and IMA for this condition
7538 check though. Patches welcome!
7540 * A new configuration file /etc/systemd/sleep.conf has been
7541 added that may be used to configure which kernel operation
7542 systemd is supposed to execute when "suspend", "hibernate"
7543 or "hybrid-sleep" is requested. This makes the new kernel
7544 "freeze" state accessible to the user.
7546 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
7547 the passed argument if applicable.
7549 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
7550 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
7551 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
7552 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
7553 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
7554 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
7555 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
7560 * The output of 'systemctl list-jobs' got some polishing. The
7561 '--type=' argument may now be passed more than once. A new
7562 command 'systemctl list-sockets' has been added which shows
7563 a list of kernel sockets systemd is listening on with the
7564 socket units they belong to, plus the units these socket
7567 * The experimental libsystemd-bus library got substantial
7568 updates to work in conjunction with the (also experimental)
7569 kdbus kernel project. It works well enough to exchange
7570 messages with some sophistication. Note that kdbus is not
7571 ready yet, and the library is mostly an elaborate test case
7572 for now, and not installable.
7574 * systemd gained a new unit 'systemd-static-nodes.service'
7575 that generates static device nodes earlier during boot, and
7576 can run in conjunction with udev.
7578 * libsystemd-login gained a new call sd_pid_get_user_unit()
7579 to retrieve the user systemd unit a process is running
7580 in. This is useful for systems where systemd is used as
7583 * systemd-nspawn now places all containers in the new /machine
7584 top-level cgroup directory in the name=systemd
7585 hierarchy. libvirt will soon do the same, so that we get a
7586 uniform separation of /system, /user and /machine for system
7587 services, user processes and containers/virtual
7588 machines. This new cgroup hierarchy is also useful to stick
7589 stable names to specific container instances, which can be
7590 recognized later this way (this name may be controlled
7591 via systemd-nspawn's new -M switch). libsystemd-login also
7592 gained a new call sd_pid_get_machine_name() to retrieve the
7593 name of the container/VM a specific process belongs to.
7595 * bootchart can now store its data in the journal.
7597 * libsystemd-journal gained a new call
7598 sd_journal_add_conjunction() for AND expressions to the
7599 matching logic. This can be used to express more complex
7600 logical expressions.
7602 * journactl can now take multiple --unit= and --user-unit=
7605 * The cryptsetup logic now understands the "luks.key=" kernel
7606 command line switch for specifying a file to read the
7607 decryption key from. Also, if a configured key file is not
7608 found the tool will now automatically fall back to prompting
7611 * Python systemd.journal module was updated to wrap recently
7612 added functions from libsystemd-journal. The interface was
7613 changed to bring the low level interface in s.j._Reader
7614 closer to the C API, and the high level interface in
7615 s.j.Reader was updated to wrap and convert all data about
7618 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
7619 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
7620 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
7621 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
7622 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
7623 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7627 * journalctl --update-catalog now understands a new --root=
7628 option to operate on catalogs found in a different root
7631 * During shutdown after systemd has terminated all running
7632 services a final killing loop kills all remaining left-over
7633 processes. We will now print the name of these processes
7634 when we send SIGKILL to them, since this usually indicates a
7637 * If /etc/crypttab refers to password files stored on
7638 configured mount points automatic dependencies will now be
7639 generated to ensure the specific mount is established first
7640 before the key file is attempted to be read.
7642 * 'systemctl status' will now show information about the
7643 network sockets a socket unit is listening on.
7645 * 'systemctl status' will also shown information about any
7646 drop-in configuration file for units. (Drop-In configuration
7647 files in this context are files such as
7648 /etc/systemd/systemd/foobar.service.d/*.conf)
7650 * systemd-cgtop now optionally shows summed up CPU times of
7651 cgroups. Press '%' while running cgtop to switch between
7652 percentage and absolute mode. This is useful to determine
7653 which cgroups use up the most CPU time over the entire
7654 runtime of the system. systemd-cgtop has also been updated
7655 to be 'pipeable' for processing with further shell tools.
7657 * 'hostnamectl set-hostname' will now allow setting of FQDN
7660 * The formatting and parsing of time span values has been
7661 changed. The parser now understands fractional expressions
7662 such as "5.5h". The formatter will now output fractional
7663 expressions for all time spans under 1min, i.e. "5.123456s"
7664 rather than "5s 123ms 456us". For time spans under 1s
7665 millisecond values are shown, for those under 1ms
7666 microsecond values are shown. This should greatly improve
7667 all time-related output of systemd.
7669 * libsystemd-login and libsystemd-journal gained new
7670 functions for querying the poll() events mask and poll()
7671 timeout value for integration into arbitrary event
7674 * localectl gained the ability to list available X11 keymaps
7675 (models, layouts, variants, options).
7677 * 'systemd-analyze dot' gained the ability to filter for
7678 specific units via shell-style globs, to create smaller,
7679 more useful graphs. I.e. it is now possible to create simple
7680 graphs of all the dependencies between only target units, or
7681 of all units that Avahi has dependencies with.
7683 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
7684 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
7685 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
7686 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
7687 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
7688 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
7689 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
7693 * The boot-time readahead implementation for rotating media
7694 will now read the read-ahead data in multiple passes which
7695 consist of all read requests made in equidistant time
7696 intervals. This means instead of strictly reading read-ahead
7697 data in its physical order on disk we now try to find a
7698 middle ground between physical and access time order.
7700 * /etc/os-release files gained a new BUILD_ID= field for usage
7701 on operating systems that provide continuous builds of OS
7704 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
7705 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
7706 William Douglas, Zbigniew Jędrzejewski-Szmek
7710 * systemd-python gained an API exposing libsystemd-daemon.
7712 * The SMACK setup logic gained support for uploading CIPSO
7715 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
7716 ReadOnlyDirectories= and InaccessibleDirectories= has
7717 changed. The private /tmp and /var/tmp directories are now
7718 shared by all processes of a service (which means
7719 ExecStartPre= may now leave data in /tmp that ExecStart= of
7720 the same service can still access). When a service is
7721 stopped its temporary directories are immediately deleted
7722 (normal clean-up with tmpfiles is still done in addition to
7725 * By default, systemd will now set a couple of sysctl
7726 variables in the kernel: the safe sysrq options are turned
7727 on, IP route verification is turned on, and source routing
7728 disabled. The recently added hardlink and softlink
7729 protection of the kernel is turned on. These settings should
7730 be reasonably safe, and good defaults for all new systems.
7732 * The predictable network naming logic may now be turned off
7733 with a new kernel command line switch: net.ifnames=0.
7735 * A new libsystemd-bus module has been added that implements a
7736 pretty complete D-Bus client library. For details see:
7738 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
7740 * journald will now explicitly flush the journal files to disk
7741 at the latest 5min after each write. The file will then also
7742 be marked offline until the next write. This should increase
7743 reliability in case of a crash. The synchronization delay
7744 can be configured via SyncIntervalSec= in journald.conf.
7746 * There's a new remote-fs-setup.target unit that can be used
7747 to pull in specific services when at least one remote file
7748 system is to be mounted.
7750 * There are new targets timers.target and paths.target as
7751 canonical targets to pull user timer and path units in
7752 from. This complements sockets.target with a similar
7753 purpose for socket units.
7755 * libudev gained a new call udev_device_set_attribute_value()
7756 to set sysfs attributes of a device.
7758 * The udev daemon now sets the default number of worker
7759 processes executed in parallel based on the number of available
7760 CPUs instead of the amount of available RAM. This is supposed
7761 to provide a more reliable default and limit a too aggressive
7762 parallelism for setups with 1000s of devices connected.
7764 Contributions from: Auke Kok, Colin Walters, Cristian
7765 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
7766 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
7767 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
7768 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
7769 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
7770 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
7771 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
7772 Zbigniew Jędrzejewski-Szmek
7776 * Configuration of unit files may now be extended via drop-in
7777 files without having to edit/override the unit files
7778 themselves. More specifically, if the administrator wants to
7779 change one value for a service file foobar.service he can
7780 now do so by dropping in a configuration snippet into
7781 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
7782 will load all these snippets and apply them on top of the
7783 main unit configuration file, possibly extending or
7784 overriding its settings. Using these drop-in snippets is
7785 generally nicer than the two earlier options for changing
7786 unit files locally: copying the files from
7787 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
7788 them there; or creating a new file in /etc/systemd/system/
7789 that incorporates the original one via ".include". Drop-in
7790 snippets into these .d/ directories can be placed in any
7791 directory systemd looks for units in, and the usual
7792 overriding semantics between /usr/lib, /etc and /run apply
7795 * Most unit file settings which take lists of items can now be
7796 reset by assigning the empty string to them. For example,
7797 normally, settings such as Environment=FOO=BAR append a new
7798 environment variable assignment to the environment block,
7799 each time they are used. By assigning Environment= the empty
7800 string the environment block can be reset to empty. This is
7801 particularly useful with the .d/*.conf drop-in snippets
7802 mentioned above, since this adds the ability to reset list
7803 settings from vendor unit files via these drop-ins.
7805 * systemctl gained a new "list-dependencies" command for
7806 listing the dependencies of a unit recursively.
7808 * Inhibitors are now honored and listed by "systemctl
7809 suspend", "systemctl poweroff" (and similar) too, not only
7810 GNOME. These commands will also list active sessions by
7813 * Resource limits (as exposed by the various control group
7814 controllers) can now be controlled dynamically at runtime
7815 for all units. More specifically, you can now use a command
7816 like "systemctl set-cgroup-attr foobar.service cpu.shares
7817 2000" to alter the CPU shares a specific service gets. These
7818 settings are stored persistently on disk, and thus allow the
7819 administrator to easily adjust the resource usage of
7820 services with a few simple commands. This dynamic resource
7821 management logic is also available to other programs via the
7822 bus. Almost any kernel cgroup attribute and controller is
7825 * systemd-vconsole-setup will now copy all font settings to
7826 all allocated VTs, where it previously applied them only to
7829 * libsystemd-login gained the new sd_session_get_tty() API
7832 * This release drops support for a few legacy or
7833 distribution-specific LSB facility names when parsing init
7834 scripts: $x-display-manager, $mail-transfer-agent,
7835 $mail-transport-agent, $mail-transfer-agent, $smtp,
7836 $null. Also, the mail-transfer-agent.target unit backing
7837 this has been removed. Distributions which want to retain
7838 compatibility with this should carry the burden for
7839 supporting this themselves and patch support for these back
7840 in, if they really need to. Also, the facilities $syslog and
7841 $local_fs are now ignored, since systemd does not support
7842 early-boot LSB init scripts anymore, and these facilities
7843 are implied anyway for normal services. syslog.target has
7846 * There are new bus calls on PID1's Manager object for
7847 cancelling jobs, and removing snapshot units. Previously,
7848 both calls were only available on the Job and Snapshot
7851 * systemd-journal-gatewayd gained SSL support.
7853 * The various "environment" files, such as /etc/locale.conf
7854 now support continuation lines with a backslash ("\") as
7855 last character in the line, similarly in style (but different)
7856 to how this is supported in shells.
7858 * For normal user processes the _SYSTEMD_USER_UNIT= field is
7859 now implicitly appended to every log entry logged. systemctl
7860 has been updated to filter by this field when operating on a
7861 user systemd instance.
7863 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
7864 CAP_AUDIT_CONTROL capabilities to the capabilities set for
7865 the container. This makes it easier to boot unmodified
7866 Fedora systems in a container, which however still requires
7867 audit=0 to be passed on the kernel command line. Auditing in
7868 kernel and userspace is unfortunately still too broken in
7869 context of containers, hence we recommend compiling it out
7870 of the kernel or using audit=0. Hopefully this will be fixed
7871 one day for good in the kernel.
7873 * nspawn gained the new --bind= and --bind-ro= parameters to
7874 bind mount specific directories from the host into the
7877 * nspawn will now mount its own devpts file system instance
7878 into the container, in order not to leak pty devices from
7879 the host into the container.
7881 * systemd will now read the firmware boot time performance
7882 information from the EFI variables, if the used boot loader
7883 supports this, and takes it into account for boot performance
7884 analysis via "systemd-analyze". This is currently supported
7885 only in conjunction with Gummiboot, but could be supported
7886 by other boot loaders too. For details see:
7888 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
7890 * A new generator has been added that automatically mounts the
7891 EFI System Partition (ESP) to /boot, if that directory
7892 exists, is empty, and no other file system has been
7893 configured to be mounted there.
7895 * logind will now send out PrepareForSleep(false) out
7896 unconditionally, after coming back from suspend. This may be
7897 used by applications as asynchronous notification for
7898 system resume events.
7900 * "systemctl unlock-sessions" has been added, that allows
7901 unlocking the screens of all user sessions at once, similar
7902 to how "systemctl lock-sessions" already locked all users
7903 sessions. This is backed by a new D-Bus call UnlockSessions().
7905 * "loginctl seat-status" will now show the master device of a
7906 seat. (i.e. the device of a seat that needs to be around for
7907 the seat to be considered available, usually the graphics
7910 * tmpfiles gained a new "X" line type, that allows
7911 configuration of files and directories (with wildcards) that
7912 shall be excluded from automatic cleanup ("aging").
7914 * udev default rules set the device node permissions now only
7915 at "add" events, and do not change them any longer with a
7916 later "change" event.
7918 * The log messages for lid events and power/sleep keypresses
7919 now carry a message ID.
7921 * We now have a substantially larger unit test suite, but this
7922 continues to be work in progress.
7924 * udevadm hwdb gained a new --root= parameter to change the
7925 root directory to operate relative to.
7927 * logind will now issue a background sync() request to the kernel
7928 early at shutdown, so that dirty buffers are flushed to disk early
7929 instead of at the last moment, in order to optimize shutdown
7932 * A new bootctl tool has been added that is an interface for
7933 certain boot loader operations. This is currently a preview
7934 and is likely to be extended into a small mechanism daemon
7935 like timedated, localed, hostnamed, and can be used by
7936 graphical UIs to enumerate available boot options, and
7937 request boot into firmware operations.
7939 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7940 the rest of the package. It also has been updated to work
7941 correctly in initrds.
7943 * polkit previously has been runtime optional, and is now also
7944 compile time optional via a configure switch.
7946 * systemd-analyze has been reimplemented in C. Also "systemctl
7947 dot" has moved into systemd-analyze.
7949 * "systemctl status" with no further parameters will now print
7950 the status of all active or failed units.
7952 * Operations such as "systemctl start" can now be executed
7953 with a new mode "--irreversible" which may be used to queue
7954 operations that cannot accidentally be reversed by a later
7955 job queuing. This is by default used to make shutdown
7956 requests more robust.
7958 * The Python API of systemd now gained a new module for
7959 reading journal files.
7961 * A new tool kernel-install has been added that can install
7962 kernel images according to the Boot Loader Specification:
7964 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
7966 * Boot time console output has been improved to provide
7967 animated boot time output for hanging jobs.
7969 * A new tool systemd-activate has been added which can be used
7970 to test socket activation with, directly from the command
7971 line. This should make it much easier to test and debug
7972 socket activation in daemons.
7974 * journalctl gained a new "--reverse" (or -r) option to show
7975 journal output in reverse order (i.e. newest line first).
7977 * journalctl gained a new "--pager-end" (or -e) option to jump
7978 to immediately jump to the end of the journal in the
7979 pager. This is only supported in conjunction with "less".
7981 * journalctl gained a new "--user-unit=" option, that works
7982 similarly to "--unit=" but filters for user units rather than
7985 * A number of unit files to ease adoption of systemd in
7986 initrds has been added. This moves some minimal logic from
7987 the various initrd implementations into systemd proper.
7989 * The journal files are now owned by a new group
7990 "systemd-journal", which exists specifically to allow access
7991 to the journal, and nothing else. Previously, we used the
7992 "adm" group for that, which however possibly covers more
7993 than just journal/log file access. This new group is now
7994 already used by systemd-journal-gatewayd to ensure this
7995 daemon gets access to the journal files and as little else
7996 as possible. Note that "make install" will also set FS ACLs
7997 up for /var/log/journal to give "adm" and "wheel" read
7998 access to it, in addition to "systemd-journal" which owns
7999 the journal files. We recommend that packaging scripts also
8000 add read access to "adm" + "wheel" to /var/log/journal, and
8001 all existing/future journal files. To normal users and
8002 administrators little changes, however packagers need to
8003 ensure to create the "systemd-journal" system group at
8004 package installation time.
8006 * The systemd-journal-gatewayd now runs as unprivileged user
8007 systemd-journal-gateway:systemd-journal-gateway. Packaging
8008 scripts need to create these system user/group at
8011 * timedated now exposes a new boolean property CanNTP that
8012 indicates whether a local NTP service is available or not.
8014 * systemd-detect-virt will now also detect xen PVs
8016 * The pstore file system is now mounted by default, if it is
8019 * In addition to the SELinux and IMA policies we will now also
8020 load SMACK policies at early boot.
8022 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
8023 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
8024 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
8025 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
8026 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
8027 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
8028 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
8029 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
8030 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
8031 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
8032 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
8033 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
8034 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
8035 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
8039 * Timer units now support calendar time events in addition to
8040 monotonic time events. That means you can now trigger a unit
8041 based on a calendar time specification such as "Thu,Fri
8042 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
8043 or fifth day of any month of the year 2013, given that it is
8044 a thursday or friday. This brings timer event support
8045 considerably closer to cron's capabilities. For details on
8046 the supported calendar time specification language see
8049 * udev now supports a number of different naming policies for
8050 network interfaces for predictable names, and a combination
8051 of these policies is now the default. Please see this wiki
8052 document for details:
8054 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8056 * Auke Kok's bootchart implementation has been added to the
8057 systemd tree. It is an optional component that can graph the
8058 boot in quite some detail. It is one of the best bootchart
8059 implementations around and minimal in its code and
8062 * nss-myhostname has been integrated into the systemd source
8063 tree. nss-myhostname guarantees that the local hostname
8064 always stays resolvable via NSS. It has been a weak
8065 requirement of systemd-hostnamed since a long time, and
8066 since its code is actually trivial we decided to just
8067 include it in systemd's source tree. It can be turned off
8068 with a configure switch.
8070 * The read-ahead logic is now capable of properly detecting
8071 whether a btrfs file system is on SSD or rotating media, in
8072 order to optimize the read-ahead scheme. Previously, it was
8073 only capable of detecting this on traditional file systems
8076 * In udev, additional device properties are now read from the
8077 IAB in addition to the OUI database. Also, Bluetooth company
8078 identities are attached to the devices as well.
8080 * In service files %U may be used as specifier that is
8081 replaced by the configured user name of the service.
8083 * nspawn may now be invoked without a controlling TTY. This
8084 makes it suitable for invocation as its own service. This
8085 may be used to set up a simple containerized server system
8086 using only core OS tools.
8088 * systemd and nspawn can now accept socket file descriptors
8089 when they are started for socket activation. This enables
8090 implementation of socket activated nspawn
8091 containers. i.e. think about autospawning an entire OS image
8092 when the first SSH or HTTP connection is received. We expect
8093 that similar functionality will also be added to libvirt-lxc
8096 * journalctl will now suppress ANSI color codes when
8097 presenting log data.
8099 * systemctl will no longer show control group information for
8100 a unit if the control group is empty anyway.
8102 * logind can now automatically suspend/hibernate/shutdown the
8105 * /etc/machine-info and hostnamed now also expose the chassis
8106 type of the system. This can be used to determine whether
8107 the local system is a laptop, desktop, handset or
8108 tablet. This information may either be configured by the
8109 user/vendor or is automatically determined from ACPI and DMI
8110 information if possible.
8112 * A number of polkit actions are now bound together with "imply"
8113 rules. This should simplify creating UIs because many actions
8114 will now authenticate similar ones as well.
8116 * Unit files learnt a new condition ConditionACPower= which
8117 may be used to conditionalize a unit depending on whether an
8118 AC power source is connected or not, of whether the system
8119 is running on battery power.
8121 * systemctl gained a new "is-failed" verb that may be used in
8122 shell scripts and suchlike to check whether a specific unit
8123 is in the "failed" state.
8125 * The EnvironmentFile= setting in unit files now supports file
8126 globbing, and can hence be used to easily read a number of
8127 environment files at once.
8129 * systemd will no longer detect and recognize specific
8130 distributions. All distribution-specific #ifdeffery has been
8131 removed, systemd is now fully generic and
8132 distribution-agnostic. Effectively, not too much is lost as
8133 a lot of the code is still accessible via explicit configure
8134 switches. However, support for some distribution specific
8135 legacy configuration file formats has been dropped. We
8136 recommend distributions to simply adopt the configuration
8137 files everybody else uses now and convert the old
8138 configuration from packaging scripts. Most distributions
8139 already did that. If that's not possible or desirable,
8140 distributions are welcome to forward port the specific
8141 pieces of code locally from the git history.
8143 * When logging a message about a unit systemd will now always
8144 log the unit name in the message meta data.
8146 * localectl will now also discover system locale data that is
8147 not stored in locale archives, but directly unpacked.
8149 * logind will no longer unconditionally use framebuffer
8150 devices as seat masters, i.e. as devices that are required
8151 to be existing before a seat is considered preset. Instead,
8152 it will now look for all devices that are tagged as
8153 "seat-master" in udev. By default, framebuffer devices will
8154 be marked as such, but depending on local systems, other
8155 devices might be marked as well. This may be used to
8156 integrate graphics cards using closed source drivers (such
8157 as NVidia ones) more nicely into logind. Note however, that
8158 we recommend using the open source NVidia drivers instead,
8159 and no udev rules for the closed-source drivers will be
8160 shipped from us upstream.
8162 Contributions from: Adam Williamson, Alessandro Crismani, Auke
8163 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
8164 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
8165 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
8166 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
8167 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
8168 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
8169 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
8170 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
8171 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
8172 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
8173 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
8178 * udev gained support for loading additional device properties
8179 from an indexed database that is keyed by vendor/product IDs
8180 and similar device identifiers. For the beginning this
8181 "hwdb" is populated with data from the well-known PCI and
8182 USB database, but also includes PNP, ACPI and OID data. In
8183 the longer run this indexed database shall grow into
8184 becoming the one central database for non-essential
8185 userspace device metadata. Previously, data from the PCI/USB
8186 database was only attached to select devices, since the
8187 lookup was a relatively expensive operation due to O(n) time
8188 complexity (with n being the number of entries in the
8189 database). Since this is now O(1), we decided to add in this
8190 data for all devices where this is available, by
8191 default. Note that the indexed database needs to be rebuilt
8192 when new data files are installed. To achieve this you need
8193 to update your packaging scripts to invoke "udevadm hwdb
8194 --update" after installation of hwdb data files. For
8195 RPM-based distributions we introduced the new
8196 %udev_hwdb_update macro for this purpose.
8198 * The Journal gained support for the "Message Catalog", an
8199 indexed database to link up additional information with
8200 journal entries. For further details please check:
8202 https://www.freedesktop.org/wiki/Software/systemd/catalog
8204 The indexed message catalog database also needs to be
8205 rebuilt after installation of message catalog files. Use
8206 "journalctl --update-catalog" for this. For RPM-based
8207 distributions we introduced the %journal_catalog_update
8208 macro for this purpose.
8210 * The Python Journal bindings gained support for the standard
8211 Python logging framework.
8213 * The Journal API gained new functions for checking whether
8214 the underlying file system of a journal file is capable of
8215 properly reporting file change notifications, or whether
8216 applications that want to reflect journal changes "live"
8217 need to recheck journal files continuously in appropriate
8220 * It is now possible to set the "age" field for tmpfiles
8221 entries to 0, indicating that files matching this entry
8222 shall always be removed when the directories are cleaned up.
8224 * coredumpctl gained a new "gdb" verb which invokes gdb
8225 right-away on the selected coredump.
8227 * There's now support for "hybrid sleep" on kernels that
8228 support this, in addition to "suspend" and "hibernate". Use
8229 "systemctl hybrid-sleep" to make use of this.
8231 * logind's HandleSuspendKey= setting (and related settings)
8232 now gained support for a new "lock" setting to simply
8233 request the screen lock on all local sessions, instead of
8234 actually executing a suspend or hibernation.
8236 * systemd will now mount the EFI variables file system by
8239 * Socket units now gained support for configuration of the
8240 SMACK security label.
8242 * timedatectl will now output the time of the last and next
8243 daylight saving change.
8245 * We dropped support for various legacy and distro-specific
8246 concepts, such as insserv, early-boot SysV services
8247 (i.e. those for non-standard runlevels such as 'b' or 'S')
8248 or ArchLinux /etc/rc.conf support. We recommend the
8249 distributions who still need support this to either continue
8250 to maintain the necessary patches downstream, or find a
8251 different solution. (Talk to us if you have questions!)
8253 * Various systemd components will now bypass polkit checks for
8254 root and otherwise handle properly if polkit is not found to
8255 be around. This should fix most issues for polkit-less
8256 systems. Quite frankly this should have been this way since
8257 day one. It is absolutely our intention to make systemd work
8258 fine on polkit-less systems, and we consider it a bug if
8259 something does not work as it should if polkit is not around.
8261 * For embedded systems it is now possible to build udev and
8262 systemd without blkid and/or kmod support.
8264 * "systemctl switch-root" is now capable of switching root
8265 more than once. I.e. in addition to transitions from the
8266 initrd to the host OS it is now possible to transition to
8267 further OS images from the host. This is useful to implement
8268 offline updating tools.
8270 * Various other additions have been made to the RPM macros
8271 shipped with systemd. Use %udev_rules_update() after
8272 installing new udev rules files. %_udevhwdbdir,
8273 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
8274 %_sysctldir are now available which resolve to the right
8275 directories for packages to place various data files in.
8277 * journalctl gained the new --full switch (in addition to
8278 --all, to disable ellipsation for long messages.
8280 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
8281 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
8282 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
8283 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
8284 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
8285 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
8286 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
8287 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
8288 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
8292 * journalctl gained new --since= and --until= switches to
8293 filter by time. It also now supports nice filtering for
8294 units via --unit=/-u.
8296 * Type=oneshot services may use ExecReload= and do the
8299 * The journal daemon now supports time-based rotation and
8300 vacuuming, in addition to the usual disk-space based
8303 * The journal will now index the available field values for
8304 each field name. This enables clients to show pretty drop
8305 downs of available match values when filtering. The bash
8306 completion of journalctl has been updated
8307 accordingly. journalctl gained a new switch -F to list all
8308 values a certain field takes in the journal database.
8310 * More service events are now written as structured messages
8311 to the journal, and made recognizable via message IDs.
8313 * The timedated, localed and hostnamed mini-services which
8314 previously only provided support for changing time, locale
8315 and hostname settings from graphical DEs such as GNOME now
8316 also have a minimal (but very useful) text-based client
8317 utility each. This is probably the nicest way to changing
8318 these settings from the command line now, especially since
8319 it lists available options and is fully integrated with bash
8322 * There's now a new tool "systemd-coredumpctl" to list and
8323 extract coredumps from the journal.
8325 * We now install a README each in /var/log/ and
8326 /etc/rc.d/init.d explaining where the system logs and init
8327 scripts went. This hopefully should help folks who go to
8328 that dirs and look into the otherwise now empty void and
8329 scratch their heads.
8331 * When user-services are invoked (by systemd --user) the
8332 $MANAGERPID env var is set to the PID of systemd.
8334 * SIGRTMIN+24 when sent to a --user instance will now result
8335 in immediate termination of systemd.
8337 * gatewayd received numerous feature additions such as a
8338 "follow" mode, for live syncing and filtering.
8340 * browse.html now allows filtering and showing detailed
8341 information on specific entries. Keyboard navigation and
8342 mouse screen support has been added.
8344 * gatewayd/journalctl now supports HTML5/JSON
8345 Server-Sent-Events as output.
8347 * The SysV init script compatibility logic will now
8348 heuristically determine whether a script supports the
8349 "reload" verb, and only then make this available as
8352 * "systemctl status --follow" has been removed, use "journalctl
8355 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
8356 have been removed since they are hardly useful to be
8359 * And I'd like to take the opportunity to specifically mention
8360 Zbigniew for his great contributions. Zbigniew, you rock!
8362 Contributions from: Andrew Eikum, Christian Hesse, Colin
8363 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
8364 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
8365 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
8366 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
8367 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
8368 Jędrzejewski-Szmek, Сковорода Никита Андреевич
8372 * If /etc/vconsole.conf is non-existent or empty we will no
8373 longer load any console font or key map at boot by
8374 default. Instead the kernel defaults will be left
8375 intact. This is definitely the right thing to do, as no
8376 configuration should mean no configuration, and hard-coding
8377 font names that are different on all archs is probably a bad
8378 idea. Also, the kernel default key map and font should be
8379 good enough for most cases anyway, and mostly identical to
8380 the userspace fonts/key maps we previously overloaded them
8381 with. If distributions want to continue to default to a
8382 non-kernel font or key map they should ship a default
8383 /etc/vconsole.conf with the appropriate contents.
8385 Contributions from: Colin Walters, Daniel J Walsh, Dave
8386 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
8387 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8391 * journalctl gained a new --cursor= switch to show entries
8392 starting from the specified location in the journal.
8394 * We now enforce a size limit on journal entry fields exported
8395 with "-o json" in journalctl. Fields larger than 4K will be
8396 assigned null. This can be turned off with --all.
8398 * An (optional) journal gateway daemon is now available as
8399 "systemd-journal-gatewayd.service". This service provides
8400 access to the journal via HTTP and JSON. This functionality
8401 will be used to implement live log synchronization in both
8402 pull and push modes, but has various other users too, such
8403 as easy log access for debugging of embedded devices. Right
8404 now it is already useful to retrieve the journal via HTTP:
8406 # systemctl start systemd-journal-gatewayd.service
8407 # wget http://localhost:19531/entries
8409 This will download the journal contents in a
8410 /var/log/messages compatible format. The same as JSON:
8412 # curl -H"Accept: application/json" http://localhost:19531/entries
8414 This service is also accessible via a web browser where a
8415 single static HTML5 app is served that uses the JSON logic
8416 to enable the user to do some basic browsing of the
8417 journal. This will be extended later on. Here's an example
8418 screenshot of this app in its current state:
8420 http://0pointer.de/public/journal-gatewayd
8422 Contributions from: Kay Sievers, Lennart Poettering, Robert
8423 Milasan, Tom Gundersen
8427 * The bash completion logic is now available for journalctl
8430 * We do not mount the "cpuset" controller anymore together with
8431 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
8432 started if no parameters are assigned to it. "cpuset" hence
8433 broke code that assumed it could create "cpu" groups and
8436 * journalctl -f will now subscribe to terminal size changes,
8437 and line break accordingly.
8439 Contributions from: Dave Reisner, Kay Sievers, Lennart
8440 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
8444 * nspawn will now create a symlink /etc/localtime in the
8445 container environment, copying the host's timezone
8446 setting. Previously this has been done via a bind mount, but
8447 since symlinks cannot be bind mounted this has now been
8448 changed to create/update the appropriate symlink.
8450 * journalctl -n's line number argument is now optional, and
8451 will default to 10 if omitted.
8453 * journald will now log the maximum size the journal files may
8454 take up on disk. This is particularly useful if the default
8455 built-in logic of determining this parameter from the file
8456 system size is used. Use "systemctl status
8457 systemd-journald.service" to see this information.
8459 * The multi-seat X wrapper tool has been stripped down. As X
8460 is now capable of enumerating graphics devices via udev in a
8461 seat-aware way the wrapper is not strictly necessary
8462 anymore. A stripped down temporary stop-gap is still shipped
8463 until the upstream display managers have been updated to
8464 fully support the new X logic. Expect this wrapper to be
8465 removed entirely in one of the next releases.
8467 * HandleSleepKey= in logind.conf has been split up into
8468 HandleSuspendKey= and HandleHibernateKey=. The old setting
8469 is not available anymore. X11 and the kernel are
8470 distinguishing between these keys and we should too. This
8471 also means the inhibition lock for these keys has been split
8474 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
8475 Poettering, Lukas Nykryn, Václav Pavlín
8479 * Whenever a unit changes state we will now log this to the
8480 journal and show along the unit's own log output in
8483 * ConditionPathIsMountPoint= can now properly detect bind
8484 mount points too. (Previously, a bind mount of one file
8485 system to another place in the same file system could not be
8486 detected as mount, since they shared struct stat's st_dev
8489 * We will now mount the cgroup controllers cpu, cpuacct,
8490 cpuset and the controllers net_cls, net_prio together by
8493 * nspawn containers will now have a virtualized boot
8494 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
8495 over with a randomized ID at container initialization). This
8496 has the effect of making "journalctl -b" do the right thing
8499 * The JSON output journal serialization has been updated not
8500 to generate "endless" list objects anymore, but rather one
8501 JSON object per line. This is more in line how most JSON
8502 parsers expect JSON objects. The new output mode
8503 "json-pretty" has been added to provide similar output, but
8504 neatly aligned for readability by humans.
8506 * We dropped all explicit sync() invocations in the shutdown
8507 code. The kernel does this implicitly anyway in the kernel
8508 reboot() syscall. halt(8)'s -n option is now a compatibility
8511 * We now support virtualized reboot() in containers, as
8512 supported by newer kernels. We will fall back to exit() if
8513 CAP_SYS_REBOOT is not available to the container. Also,
8514 nspawn makes use of this now and will actually reboot the
8515 container if the containerized OS asks for that.
8517 * journalctl will only show local log output by default
8518 now. Use --merge (-m) to show remote log output, too.
8520 * libsystemd-journal gained the new sd_journal_get_usage()
8521 call to determine the current disk usage of all journal
8522 files. This is exposed in the new "journalctl --disk-usage"
8525 * journald gained a new configuration setting SplitMode= in
8526 journald.conf which may be used to control how user journals
8527 are split off. See journald.conf(5) for details.
8529 * A new condition type ConditionFileNotEmpty= has been added.
8531 * tmpfiles' "w" lines now support file globbing, to write
8532 multiple files at once.
8534 * We added Python bindings for the journal submission
8535 APIs. More Python APIs for a number of selected APIs will
8536 likely follow. Note that we intend to add native bindings
8537 only for the Python language, as we consider it common
8538 enough to deserve bindings shipped within systemd. There are
8539 various projects outside of systemd that provide bindings
8540 for languages such as PHP or Lua.
8542 * Many conditions will now resolve specifiers such as %i. In
8543 addition, PathChanged= and related directives of .path units
8544 now support specifiers as well.
8546 * There's now a new RPM macro definition for the system preset
8549 * journald will now warn if it ca not forward a message to the
8550 syslog daemon because its socket is full.
8552 * timedated will no longer write or process /etc/timezone,
8553 except on Debian. As we do not support late mounted /usr
8554 anymore /etc/localtime always being a symlink is now safe,
8555 and hence the information in /etc/timezone is not necessary
8558 * logind will now always reserve one VT for a text getty (VT6
8559 by default). Previously if more than 6 X sessions where
8560 started they took up all the VTs with auto-spawned gettys,
8561 so that no text gettys were available anymore.
8563 * udev will now automatically inform the btrfs kernel logic
8564 about btrfs RAID components showing up. This should make
8565 simple hotplug based btrfs RAID assembly work.
8567 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
8568 (but not for its children which will stay at the kernel
8569 default). This should allow setups with a lot more listening
8572 * systemd will now always pass the configured timezone to the
8573 kernel at boot. timedated will do the same when the timezone
8576 * logind's inhibition logic has been updated. By default,
8577 logind will now handle the lid switch, the power and sleep
8578 keys all the time, even in graphical sessions. If DEs want
8579 to handle these events on their own they should take the new
8580 handle-power-key, handle-sleep-key and handle-lid-switch
8581 inhibitors during their runtime. A simple way to achieve
8582 that is to invoke the DE wrapped in an invocation of:
8584 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
8586 * Access to unit operations is now checked via SELinux taking
8587 the unit file label and client process label into account.
8589 * systemd will now notify the administrator in the journal
8590 when he over-mounts a non-empty directory.
8592 * There are new specifiers that are resolved in unit files,
8593 for the host name (%H), the machine ID (%m) and the boot ID
8596 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
8597 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
8598 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
8599 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
8600 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
8601 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
8602 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
8606 * Support for reading structured kernel messages from
8607 /dev/kmsg has now been added and is enabled by default.
8609 * Support for reading kernel messages from /proc/kmsg has now
8610 been removed. If you want kernel messages in the journal
8611 make sure to run a recent kernel (>= 3.5) that supports
8612 reading structured messages from /dev/kmsg (see
8613 above). /proc/kmsg is now exclusive property of classic
8614 syslog daemons again.
8616 * The libudev API gained the new
8617 udev_device_new_from_device_id() call.
8619 * The logic for file system namespace (ReadOnlyDirectory=,
8620 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
8621 require pivot_root() anymore. This means fewer temporary
8622 directories are created below /tmp for this feature.
8624 * nspawn containers will now see and receive all submounts
8625 made on the host OS below the root file system of the
8628 * Forward Secure Sealing is now supported for Journal files,
8629 which provide cryptographical sealing of journal files so
8630 that attackers cannot alter log history anymore without this
8631 being detectable. Lennart will soon post a blog story about
8632 this explaining it in more detail.
8634 * There are two new service settings RestartPreventExitStatus=
8635 and SuccessExitStatus= which allow configuration of exit
8636 status (exit code or signal) which will be excepted from the
8637 restart logic, resp. consider successful.
8639 * journalctl gained the new --verify switch that can be used
8640 to check the integrity of the structure of journal files and
8641 (if Forward Secure Sealing is enabled) the contents of
8644 * nspawn containers will now be run with /dev/stdin, /dev/fd/
8645 and similar symlinks pre-created. This makes running shells
8646 as container init process a lot more fun.
8648 * The fstab support can now handle PARTUUID= and PARTLABEL=
8651 * A new ConditionHost= condition has been added to match
8652 against the hostname (with globs) and machine ID. This is
8653 useful for clusters where a single OS image is used to
8654 provision a large number of hosts which shall run slightly
8655 different sets of services.
8657 * Services which hit the restart limit will now be placed in a
8660 Contributions from: Bertram Poettering, Dave Reisner, Huang
8661 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
8662 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
8666 * When running in --user mode systemd will now become a
8667 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
8668 tree a lot more organized.
8670 * A new PartOf= unit dependency type has been introduced that
8671 may be used to group services in a natural way.
8673 * "systemctl enable" may now be used to enable instances of
8676 * journalctl now prints error log levels in red, and
8677 warning/notice log levels in bright white. It also supports
8678 filtering by log level now.
8680 * cgtop gained a new -n switch (similar to top), to configure
8681 the maximum number of iterations to run for. It also gained
8682 -b, to run in batch mode (accepting no input).
8684 * The suffix ".service" may now be omitted on most systemctl
8685 command lines involving service unit names.
8687 * There's a new bus call in logind to lock all sessions, as
8688 well as a loginctl verb for it "lock-sessions".
8690 * libsystemd-logind.so gained a new call sd_journal_perror()
8691 that works similar to libc perror() but logs to the journal
8692 and encodes structured information about the error number.
8694 * /etc/crypttab entries now understand the new keyfile-size=
8697 * shutdown(8) now can send a (configurable) wall message when
8698 a shutdown is cancelled.
8700 * The mount propagation mode for the root file system will now
8701 default to "shared", which is useful to make containers work
8702 nicely out-of-the-box so that they receive new mounts from
8703 the host. This can be undone locally by running "mount
8704 --make-rprivate /" if needed.
8706 * The prefdm.service file has been removed. Distributions
8707 should maintain this unit downstream if they intend to keep
8708 it around. However, we recommend writing normal unit files
8709 for display managers instead.
8711 * Since systemd is a crucial part of the OS we will now
8712 default to a number of compiler switches that improve
8713 security (hardening) such as read-only relocations, stack
8714 protection, and suchlike.
8716 * The TimeoutSec= setting for services is now split into
8717 TimeoutStartSec= and TimeoutStopSec= to allow configuration
8718 of individual time outs for the start and the stop phase of
8721 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
8722 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
8723 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
8724 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
8725 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
8726 Gundersen, Zbigniew Jędrzejewski-Szmek
8730 * The journal and id128 C APIs are now fully documented as man
8733 * Extra safety checks have been added when transitioning from
8734 the initial RAM disk to the main system to avoid accidental
8737 * /etc/crypttab entries now understand the new keyfile-offset=
8740 * systemctl -t can now be used to filter by unit load state.
8742 * The journal C API gained the new sd_journal_wait() call to
8743 make writing synchronous journal clients easier.
8745 * journalctl gained the new -D switch to show journals from a
8748 * journalctl now displays a special marker between log
8749 messages of two different boots.
8751 * The journal is now explicitly flushed to /var via a service
8752 systemd-journal-flush.service, rather than implicitly simply
8753 by seeing /var/log/journal to be writable.
8755 * journalctl (and the journal C APIs) can now match for much
8756 more complex expressions, with alternatives and
8759 * When transitioning from the initial RAM disk to the main
8760 system we will now kill all processes in a killing spree to
8761 ensure no processes stay around by accident.
8763 * Three new specifiers may be used in unit files: %u, %h, %s
8764 resolve to the user name, user home directory resp. user
8765 shell. This is useful for running systemd user instances.
8767 * We now automatically rotate journal files if their data
8768 object hash table gets a fill level > 75%. We also size the
8769 hash table based on the configured maximum file size. This
8770 together should lower hash collisions drastically and thus
8771 speed things up a bit.
8773 * journalctl gained the new "--header" switch to introspect
8774 header data of journal files.
8776 * A new setting SystemCallFilters= has been added to services
8777 which may be used to apply blacklists or whitelists to
8778 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
8780 * nspawn gained a new --link-journal= switch (and quicker: -j)
8781 to link the container journal with the host. This makes it
8782 very easy to centralize log viewing on the host for all
8783 guests while still keeping the journal files separated.
8785 * Many bugfixes and optimizations
8787 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
8788 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
8789 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
8794 * Several tools now understand kernel command line arguments,
8795 which are only read when run in an initial RAM disk. They
8796 usually follow closely their normal counterparts, but are
8799 * There's a new tool to analyze the readahead files that are
8800 automatically generated at boot. Use:
8802 /usr/lib/systemd/systemd-readahead analyze /.readahead
8804 * We now provide an early debug shell on tty9 if this enabled. Use:
8806 systemctl enable debug-shell.service
8808 * All plymouth related units have been moved into the Plymouth
8809 package. Please make sure to upgrade your Plymouth version
8812 * systemd-tmpfiles now supports getting passed the basename of
8813 a configuration file only, in which case it will look for it
8814 in all appropriate directories automatically.
8816 * udevadm info now takes a /dev or /sys path as argument, and
8817 does the right thing. Example:
8819 udevadm info /dev/sda
8820 udevadm info /sys/class/block/sda
8822 * systemctl now prints a warning if a unit is stopped but a
8823 unit that might trigger it continues to run. Example: a
8824 service is stopped but the socket that activates it is left
8827 * "systemctl status" will now mention if the log output was
8828 shortened due to rotation since a service has been started.
8830 * The journal API now exposes functions to determine the
8831 "cutoff" times due to rotation.
8833 * journald now understands SIGUSR1 and SIGUSR2 for triggering
8834 immediately flushing of runtime logs to /var if possible,
8835 resp. for triggering immediate rotation of the journal
8838 * It is now considered an error if a service is attempted to
8839 be stopped that is not loaded.
8841 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
8843 * systemd-analyze now supports Python 3
8845 * tmpfiles now supports cleaning up directories via aging
8846 where the first level dirs are always kept around but
8847 directories beneath it automatically aged. This is enabled
8848 by prefixing the age field with '~'.
8850 * Seat objects now expose CanGraphical, CanTTY properties
8851 which is required to deal with very fast bootups where the
8852 display manager might be running before the graphics drivers
8853 completed initialization.
8855 * Seat objects now expose a State property.
8857 * We now include RPM macros for service enabling/disabling
8858 based on the preset logic. We recommend RPM based
8859 distributions to make use of these macros if possible. This
8860 makes it simpler to reuse RPM spec files across
8863 * We now make sure that the collected systemd unit name is
8864 always valid when services log to the journal via
8867 * There's a new man page kernel-command-line(7) detailing all
8868 command line options we understand.
8870 * The fstab generator may now be disabled at boot by passing
8871 fstab=0 on the kernel command line.
8873 * A new kernel command line option modules-load= is now understood
8874 to load a specific kernel module statically, early at boot.
8876 * Unit names specified on the systemctl command line are now
8877 automatically escaped as needed. Also, if file system or
8878 device paths are specified they are automatically turned
8879 into the appropriate mount or device unit names. Example:
8881 systemctl status /home
8882 systemctl status /dev/sda
8884 * The SysVConsole= configuration option has been removed from
8885 system.conf parsing.
8887 * The SysV search path is no longer exported on the D-Bus
8890 * The Names= option has been removed from unit file parsing.
8892 * There's a new man page bootup(7) detailing the boot process.
8894 * Every unit and every generator we ship with systemd now
8895 comes with full documentation. The self-explanatory boot is
8898 * A couple of services gained "systemd-" prefixes in their
8899 name if they wrap systemd code, rather than only external
8900 code. Among them fsck@.service which is now
8901 systemd-fsck@.service.
8903 * The HaveWatchdog property has been removed from the D-Bus
8906 * systemd.confirm_spawn= on the kernel command line should now
8909 * There's a new man page crypttab(5) which details all options
8910 we actually understand.
8912 * systemd-nspawn gained a new --capability= switch to pass
8913 additional capabilities to the container.
8915 * timedated will now read known NTP implementation unit names
8916 from /usr/lib/systemd/ntp-units.d/*.list,
8917 systemd-timedated-ntp.target has been removed.
8919 * journalctl gained a new switch "-b" that lists log data of
8920 the current boot only.
8922 * The notify socket is in the abstract namespace again, in
8923 order to support daemons which chroot() at start-up.
8925 * There is a new Storage= configuration option for journald
8926 which allows configuration of where log data should go. This
8927 also provides a way to disable journal logging entirely, so
8928 that data collected is only forwarded to the console, the
8929 kernel log buffer or another syslog implementation.
8931 * Many bugfixes and optimizations
8933 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8934 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8935 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8936 Shawn Landden, Tom Gundersen
8940 * "systemctl help <unit>" now shows the man page if one is
8943 * Several new man pages have been added.
8945 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8946 MaxLevelConsole= can now be specified in
8947 journald.conf. These options allow reducing the amount of
8948 data stored on disk or forwarded by the log level.
8950 * TimerSlackNSec= can now be specified in system.conf for
8951 PID1. This allows system-wide power savings.
8953 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8954 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8959 * logind is now capable of (optionally) handling power and
8960 sleep keys as well as the lid switch.
8962 * journalctl now understands the syntax "journalctl
8963 /usr/bin/avahi-daemon" to get all log output of a specific
8966 * CapabilityBoundingSet= in system.conf now also influences
8967 the capability bound set of usermode helpers of the kernel.
8969 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8970 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8971 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8972 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8976 * Note that we skipped 139 releases here in order to set the
8977 new version to something that is greater than both udev's
8978 and systemd's most recent version number.
8980 * udev: all udev sources are merged into the systemd source tree now.
8981 All future udev development will happen in the systemd tree. It
8982 is still fully supported to use the udev daemon and tools without
8983 systemd running, like in initramfs or other init systems. Building
8984 udev though, will require the *build* of the systemd tree, but
8985 udev can be properly *run* without systemd.
8987 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
8988 should be used to create dead device nodes as workarounds for broken
8991 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8992 no longer supported. udev_monitor_new_from_netlink() needs to be
8993 used to subscribe to events.
8995 * udev: when udevd is started by systemd, processes which are left
8996 behind by forking them off of udev rules, are unconditionally cleaned
8997 up and killed now after the event handling has finished. Services or
8998 daemons must be started as systemd services. Services can be
8999 pulled-in by udev to get started, but they can no longer be directly
9000 forked by udev rules.
9002 * udev: the daemon binary is called systemd-udevd now and installed
9003 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
9004 to adapt to that, create symlink, or rename the binary after building
9007 * libudev no longer provides these symbols:
9008 udev_monitor_from_socket()
9009 udev_queue_get_failed_list_entry()
9010 udev_get_{dev,sys,run}_path()
9011 The versions number was bumped and symbol versioning introduced.
9013 * systemd-loginctl and systemd-journalctl have been renamed
9014 to loginctl and journalctl to match systemctl.
9016 * The config files: /etc/systemd/systemd-logind.conf and
9017 /etc/systemd/systemd-journald.conf have been renamed to
9018 logind.conf and journald.conf. Package updates should rename
9019 the files to the new names on upgrade.
9021 * For almost all files the license is now LGPL2.1+, changed
9022 from the previous GPL2.0+. Exceptions are some minor stuff
9023 of udev (which will be changed to LGPL2.1 eventually, too),
9024 and the MIT licensed sd-daemon.[ch] library that is suitable
9025 to be used as drop-in files.
9027 * systemd and logind now handle system sleep states, in
9028 particular suspending and hibernating.
9030 * logind now implements a sleep/shutdown/idle inhibiting logic
9031 suitable for a variety of uses. Soonishly Lennart will blog
9032 about this in more detail.
9034 * var-run.mount and var-lock.mount are no longer provided
9035 (which previously bind mounted these directories to their new
9036 places). Distributions which have not converted these
9037 directories to symlinks should consider stealing these files
9038 from git history and add them downstream.
9040 * We introduced the Documentation= field for units and added
9041 this to all our shipped units. This is useful to make it
9042 easier to explore the boot and the purpose of the various
9045 * All smaller setup units (such as
9046 systemd-vconsole-setup.service) now detect properly if they
9047 are run in a container and are skipped when
9048 appropriate. This guarantees an entirely noise-free boot in
9049 Linux container environments such as systemd-nspawn.
9051 * A framework for implementing offline system updates is now
9052 integrated, for details see:
9053 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
9055 * A new service type Type=idle is available now which helps us
9056 avoiding ugly interleaving of getty output and boot status
9059 * There's now a system-wide CapabilityBoundingSet= option to
9060 globally reduce the set of capabilities for the
9061 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
9062 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
9063 even CAP_NET_ADMIN system-wide for secure systems.
9065 * There are now system-wide DefaultLimitXXX= options to
9066 globally change the defaults of the various resource limits
9067 for all units started by PID 1.
9069 * Harald Hoyer's systemd test suite has been integrated into
9070 systemd which allows easy testing of systemd builds in qemu
9071 and nspawn. (This is really awesome! Ask us for details!)
9073 * The fstab parser is now implemented as generator, not inside
9076 * systemctl will now warn you if .mount units generated from
9077 /etc/fstab are out of date due to changes in fstab that
9078 have not been read by systemd yet.
9080 * systemd is now suitable for usage in initrds. Dracut has
9081 already been updated to make use of this. With this in place
9082 initrds get a slight bit faster but primarily are much
9083 easier to introspect and debug since "systemctl status" in
9084 the host system can be used to introspect initrd services,
9085 and the journal from the initrd is kept around too.
9087 * systemd-delta has been added, a tool to explore differences
9088 between user/admin configuration and vendor defaults.
9090 * PrivateTmp= now affects both /tmp and /var/tmp.
9092 * Boot time status messages are now much prettier and feature
9093 proper english language. Booting up systemd has never been
9096 * Read-ahead pack files now include the inode number of all
9097 files to pre-cache. When the inode changes the pre-caching
9098 is not attempted. This should be nicer to deal with updated
9099 packages which might result in changes of read-ahead
9102 * We now temporaritly lower the kernel's read_ahead_kb variable
9103 when collecting read-ahead data to ensure the kernel's
9104 built-in read-ahead does not add noise to our measurements
9105 of necessary blocks to pre-cache.
9107 * There's now RequiresMountsFor= to add automatic dependencies
9108 for all mounts necessary for a specific file system path.
9110 * MountAuto= and SwapAuto= have been removed from
9111 system.conf. Mounting file systems at boot has to take place
9114 * nspawn now learned a new switch --uuid= to set the machine
9115 ID on the command line.
9117 * nspawn now learned the -b switch to automatically search
9120 * vt102 is now the default TERM for serial TTYs, upgraded from
9123 * systemd-logind now works on VT-less systems.
9125 * The build tree has been reorganized. The individual
9126 components now have directories of their own.
9128 * A new condition type ConditionPathIsReadWrite= is now available.
9130 * nspawn learned the new -C switch to create cgroups for the
9131 container in other hierarchies.
9133 * We now have support for hardware watchdogs, configurable in
9136 * The scheduled shutdown logic now has a public API.
9138 * We now mount /tmp as tmpfs by default, but this can be
9139 masked and /etc/fstab can override it.
9141 * Since udisks does not make use of /media anymore we are not
9142 mounting a tmpfs on it anymore.
9144 * journalctl gained a new --local switch to only interleave
9145 locally generated journal files.
9147 * We can now load the IMA policy at boot automatically.
9149 * The GTK tools have been split off into a systemd-ui.
9151 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
9152 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
9153 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
9154 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
9155 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
9156 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
9157 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
9158 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
9159 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
9164 * This is mostly a bugfix release
9166 * Support optional initialization of the machine ID from the
9167 KVM or container configured UUID.
9169 * Support immediate reboots with "systemctl reboot -ff"
9171 * Show /etc/os-release data in systemd-analyze output
9173 * Many bugfixes for the journal, including endianness fixes and
9174 ensuring that disk space enforcement works
9176 * sd-login.h is C++ compatible again
9178 * Extend the /etc/os-release format on request of the Debian
9181 * We now refuse non-UTF8 strings used in various configuration
9182 and unit files. This is done to ensure we do not pass invalid
9183 data over D-Bus or expose it elsewhere.
9185 * Register Mimo USB Screens as suitable for automatic seat
9188 * Read SELinux client context from journal clients in a race
9191 * Reorder configuration file lookup order. /etc now always
9192 overrides /run in order to allow the administrator to always
9193 and unconditionally override vendor-supplied or
9194 automatically generated data.
9196 * The various user visible bits of the journal now have man
9197 pages. We still lack man pages for the journal API calls
9200 * We now ship all man pages in HTML format again in the
9203 Contributions from: Dave Reisner, Dirk Eibach, Frederic
9204 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
9205 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
9210 * This is mostly a bugfix release
9212 * systems lacking /etc/os-release are no longer supported.
9214 * Various functionality updates to libsystemd-login.so
9216 * Track class of PAM logins to distinguish greeters from
9219 Contributions from: Kay Sievers, Lennart Poettering, Michael
9224 * This is an important bugfix release for v41.
9226 * Building man pages is now optional which should be useful
9227 for those building systemd from git but unwilling to install
9230 * Watchdog support for supervising services is now usable. In
9231 a future release support for hardware watchdogs
9232 (i.e. /dev/watchdog) will be added building on this.
9234 * Service start rate limiting is now configurable and can be
9235 turned off per service. When a start rate limit is hit a
9236 reboot can automatically be triggered.
9238 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
9240 Contributions from: Benjamin Franzke, Bill Nottingham,
9241 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
9242 Schmidt, Michał Górny, Piotr Drąg
9246 * The systemd binary is installed /usr/lib/systemd/systemd now;
9247 An existing /sbin/init symlink needs to be adapted with the
9250 * The code that loads kernel modules has been ported to invoke
9251 libkmod directly, instead of modprobe. This means we do not
9252 support systems with module-init-tools anymore.
9254 * Watchdog support is now already useful, but still not
9257 * A new kernel command line option systemd.setenv= is
9258 understood to set system wide environment variables
9259 dynamically at boot.
9261 * We now limit the set of capabilities of systemd-journald.
9263 * We now set SIGPIPE to ignore by default, since it only is
9264 useful in shell pipelines, and has little use in general
9265 code. This can be disabled with IgnoreSIPIPE=no in unit
9268 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
9269 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
9274 * This is mostly a bugfix release
9276 * We now expose the reason why a service failed in the
9277 "Result" D-Bus property.
9279 * Rudimentary service watchdog support (will be completed over
9280 the next few releases.)
9282 * When systemd forks off in order execute some service we will
9283 now immediately changes its argv[0] to reflect which process
9284 it will execute. This is useful to minimize the time window
9285 with a generic argv[0], which makes bootcharts more useful
9287 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
9288 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
9289 Mike Kazantsev, Ray Strode
9293 * This is mostly a test release, but incorporates many
9296 * New systemd-cgtop tool to show control groups by their
9299 * Linking against libacl for ACLs is optional again. If
9300 disabled, support tracking device access for active logins
9301 goes becomes unavailable, and so does access to the user
9302 journals by the respective users.
9304 * If a group "adm" exists, journal files are automatically
9305 owned by them, thus allow members of this group full access
9306 to the system journal as well as all user journals.
9308 * The journal now stores the SELinux context of the logging
9309 client for all entries.
9311 * Add C++ inclusion guards to all public headers
9313 * New output mode "cat" in the journal to print only text
9314 messages, without any meta data like date or time.
9316 * Include tiny X server wrapper as a temporary stop-gap to
9317 teach XOrg udev display enumeration. This is used by display
9318 managers such as gdm, and will go away as soon as XOrg
9319 learned native udev hotplugging for display devices.
9321 * Add new systemd-cat tool for executing arbitrary programs
9322 with STDERR/STDOUT connected to the journal. Can also act as
9323 BSD logger replacement, and does so by default.
9325 * Optionally store all locally generated coredumps in the
9326 journal along with meta data.
9328 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
9329 writing short strings to files (for usage for /sys), and for
9330 creating symlinks, character and block device nodes.
9332 * New unit file option ControlGroupPersistent= to make cgroups
9333 persistent, following the mechanisms outlined in
9334 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
9336 * Support multiple local RTCs in a sane way
9338 * No longer monopolize IO when replaying readahead data on
9339 rotating disks, since we might starve non-file-system IO to
9340 death, since fanotify() will not see accesses done by blkid,
9343 * Do not show kernel threads in systemd-cgls anymore, unless
9344 requested with new -k switch.
9346 Contributions from: Dan Horák, Kay Sievers, Lennart
9347 Poettering, Michal Schmidt
9351 * This is mostly a test release, but incorporates many
9354 * The git repository moved to:
9355 git://anongit.freedesktop.org/systemd/systemd
9356 ssh://git.freedesktop.org/git/systemd/systemd
9358 * First release with the journal
9359 http://0pointer.de/blog/projects/the-journal.html
9361 * The journal replaces both systemd-kmsg-syslogd and
9362 systemd-stdout-bridge.
9364 * New sd_pid_get_unit() API call in libsystemd-logind
9366 * Many systemadm clean-ups
9368 * Introduce remote-fs-pre.target which is ordered before all
9369 remote mounts and may be used to start services before all
9372 * Added Mageia support
9374 * Add bash completion for systemd-loginctl
9376 * Actively monitor PID file creation for daemons which exit in
9377 the parent process before having finished writing the PID
9378 file in the daemon process. Daemons which do this need to be
9379 fixed (i.e. PID file creation must have finished before the
9380 parent exits), but we now react a bit more gracefully to them.
9382 * Add colourful boot output, mimicking the well-known output
9383 of existing distributions.
9385 * New option PassCredentials= for socket units, for
9386 compatibility with a recent kernel ABI breakage.
9388 * /etc/rc.local is now hooked in via a generator binary, and
9389 thus will no longer act as synchronization point during
9392 * systemctl list-unit-files now supports --root=.
9394 * systemd-tmpfiles now understands two new commands: z, Z for
9395 relabelling files according to the SELinux database. This is
9396 useful to apply SELinux labels to specific files in /sys,
9399 * Output of SysV services is now forwarded to both the console
9400 and the journal by default, not only just the console.
9402 * New man pages for all APIs from libsystemd-login.
9404 * The build tree got reorganized and the build system is a
9405 lot more modular allowing embedded setups to specifically
9406 select the components of systemd they are interested in.
9408 * Support for Linux systems lacking the kernel VT subsystem is
9411 * configure's --with-rootdir= got renamed to
9412 --with-rootprefix= to follow the naming used by udev and
9415 * Unless specified otherwise we will now install to /usr instead
9416 of /usr/local by default.
9418 * Processes with '@' in argv[0][0] are now excluded from the
9419 final shut-down killing spree, following the logic explained
9421 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
9423 * All processes remaining in a service cgroup when we enter
9424 the START or START_PRE states are now killed with
9425 SIGKILL. That means it is no longer possible to spawn
9426 background processes from ExecStart= lines (which was never
9427 supported anyway, and bad style).
9429 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
9430 reloading of units together.
9432 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
9433 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
9434 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
9435 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
9436 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek